
9/18 
2004/3/1920 [Computer/Theory] UID:12765 Activity:nil 
3/19 I read the link someone posted about Wolfram vs. Kahan. So maybe I'm ignorant, but it seems to assert that (1/r)^x r^x, for arbitrary r, is not 1. So, why not? Is it because of r=0? \_ yes. \_ KAHAN WINS! GENGHISALITY! \_ Fractional powers can't be defined in a single continuous way for the whole complex plane. You can define a principal value, but it screws up equations like the above. E.G. r=1, x=.5 (1)^.5 * (1)^.5 = i * i = 1 != 1 \_ Partha, would you please write an article describing WvsK? \ "go ask william what the wolfram said" ... talking to kahan is pretty much a guaranteed interesting experience. psb \_ Some of us are do not live in berkeley anymore .... \_ And that was definitely the most exciting part of the entire 20th century!! Yes, indeedy! 
2004/3/1920 [Computer/Theory] UID:12751 Activity:low 
3/18 I have a question for you theoretical CS types out there. Has Stephen Wolfram ever actually contributed anything significant to the field of cellular automata? \_ Were you listening to Berkeley Groks too? Wasn't he just describing CS? \_ I don't think so. It seems like he tried to take the concept and apply it to life, the universe, and everything, while being pompous. \_ When Wolfram was a physicist at Caltech, he did a bunch of pioneering work in CA (early '80s). \_ you must be talking about his already refuted, selfpublished tome \ stephan wolfram vs. william kahan in 10evans was one of the high point of the last century. psb \_ high points of the 20th century? okey dokey! maybe you lived in a different 20th century from the rest of us. \_ were you there at kahan wolfram 1? i can't speak to the entire 20th century, but it certainly was one of the high points of my life. was there ever a kahan wolfram 2? \_ don't get out much huh? \_ What happened? \_ Partha, for the benefit of mankind, would you describe it? I am very interested, having been a student of Kahn and a customer of Wolfram. \_ i assume he is talking about this: http://tinyurl.com/2vlae \ this misses the 1+1 = 3 part. psb which doesn't really seem like any sort of highlight except for flamemongers i guess. \_ wow! that sure is cool. and to think I was all excited about the fall of the Berlin Wall and Communism. What a fool I was for missing out on the really important historical events. \_ i take it this means you were not physically present. \_ unless someone got stabbed while screaming, "oh contraire, mon frere!" it isn't important. \_ Are you talking about A_New_Kind_Of_Science? Have you read it? 
2004/3/2 [Computer/Theory, Computer/SW/Unix] UID:29844 Activity:high 
3/1 Computer science trivia of the day: Say you have X memory and have 8X that size to sort on the disk. Each time you swap from X memory to one of the 8 pages, it will cost you n time. Using what you know about merge sort, quick sort, and what not, what is the complexity in terms of n? \_ Do your own homework. \_ I used to know that but then I graduated and stopped spending my time on useless homework problems. It was O of something. 
2004/2/15 [Computer/Theory] UID:12265 Activity:very high 
2/15 Ok we disagree on many issues, but can we agree on something? Start your list here, and delete any entry you don't like (hence an intersection of issues we agree on): bush is stupid math is fun linux is better than Winblows berkeley is better than Cal State berkeley women are uglier than Cal State chix \_ Rather than deleting, wouldn't it be more fun to start flamewars under each subject? I disagree at least somewhat with all of those except the ugly one. It probably applies to guys too though. \_ are you talking shit about math? \_ I guess some math is fun, sometimes. I'm a math centrist. \_ Mushy head! 
2004/2/3 [Computer/Theory] UID:12091 Activity:nil 66%like:10029 
2/3 Happy 02/03/04! \_ What the fuck is so happy about it? \_ If you're a numerology idiot it probably means something to you. I was chillin with the Tooth Fairy last night. She was bummed that Santa keeps telling her he's going to file soon but never does, but she met this oddly cool egg laying Bunny and went home with him so it's all for the best. \_ are you guys J0X1N?? This day is so cool... personaly, I'm counting down to 3/4/05... Pythagorean Day! wheee \_ Damn bitch. Not bj? She must've watched the Vagina Monologues. \_ 01/04/09 : Alien Monolith Day! \_ Ah man, I forgot about Fibonacci day: 01/02/03 5:08 
2004/1/28 [Computer/Theory, Academia/Berkeley/CSUA/Motd] UID:11979 Activity:nil 
1/28 Slashdot has a story about the DeCSS haiku, which reference's the CSUA's collection of zinc haiku. What is he talking about? Motd historians, your city needs you. \_ ~scotsman/pub/humor/zinc_haikus \_ Some of you people are getting carried away with the apostrophes. \_ What than would you have us do? It's better then abuse of the then/than misspellings. And corrected; sorry. \_ I hope you know you just made a then/than mistake in that post. \_ http://www.loyalty.org/~schoen/haiku.html  my article including influence of CSUA zinc haiku in my poem about DeCSS.  schoen 
2004/1/2627 [Computer/Theory] UID:11956 Activity:high 
1/26 Please list what you think are the most influential/important math/science algorithms of the 20th century: (PeterM) \_ Let's try this one again... \_ Fast Fourier Transform \_ quicksort \_ fast fourier transform \_ Simplex, resolution theorem proving, belief propagation, Shor's quantum factoring.  ilyas \_ He said 20th century, not 21st, jackass. Most poeple wouldn't regard factoring the number 15 as influential/important. I think that the factoring of 15 was done after 2000 anyway. \_ Those sound pretty obscure. \_ to someone who can't even format properly, no doubt. \_ RSA to encrypt billions of dollars of transactions business guy \_ Maybe a more general statement would be "public key crypto" \_ Not an algorithm, maybe a class of algorithms, or better, a research area. \_ Generational garbage collection \_ 2+2=5 \_ ee equals em cee squared. Boom. \_ That's an equation; I don't think that's what the OP was asking for. \_ quicksort \_ fast fourier transform \_ hamsterdance \_ Simplex, resolution theorem proving, belief propagation, Shor's quantum factoring.  ilyas \_ philcompress \_ "1) invent algorithm, 2) ..., 3) profit!" is my choice. \_ Hashes / hash tables \_ That's a data structure, not an algorithm, and it only works because we have constant time lookup up to 2^32 (or 64). You can't do better than O(log n) access. \_ Generational garbage collection \_ all answered purged because some humorless jerk censored my very harmless 1 line joke. fuck you. i'm taking your ball home with me. \_ Big man. And what a sense of humor too. 
2004/1/25 [Consumer/Camera, Computer/Companies/Google, Computer/Theory] UID:11648 Activity:nil 
1/2 Computer science question for you PHDs out there. What would be the feasibility of a program that, given a set of images that form a mosaic, produce the "most correct" composite image, based upon some definition of correctness that could be supplied in advance (color compatibility, smoothness of lines, etc)? For instance, such a mosaic could be a 360 series of photos that form a panoramic photograph. Such a program wouldn't necessarily need to be perfect, and any of amount of "hints" could be given to the program as well as the input images. This might already exist, or it might be solving the halting problem  I don't know. I'm asking the question for a materials science postdoc friend of mine that is working with crystal lattice images. lye \_ Does the camera rotate as it takes pictures? Is there overlap between individual images? At its most general, this problem involves object recognition and so is visionhard. There are some papers on this problem, google for obvious things to find them. \_ such programs exist. if the camera undergoes pure rotation and no translation about the optical center, the problem is very easy to solve (assuming overlap between the views). other wise it's harder and you have to rely on some kind of approximation because you need to know the 3D geometry of the scene. other things that help: if you know the exact motion of the camera, the problem is easy again. ali \_ There is a ton of literature on this kind of problem, which is known as "registration." A standard approach is to define some kind of error function (distance between edges, or distance between overlapping pixels in color space, etc) and try to minimize it over the space of transformations. An algorithm that works well for a lot of problems of this type in the pairwise case is called "iterative closest point," due to Besl and McKay. If you don't know an approximate solution to start with, it is a lot more difficult. lewis \_ homeslide, iterative closest point requires you to know the 3D geometry of the scene to perform registration. altneratively, you need some kind of parametric transformation model for your images. 
2003/12/4 [Science/Physics, Computer/Theory] UID:11308 Activity:nil 
12/3 Did anybody catch the Nova series on PBS a month or so ago on String Theory? What do you think of it? Is this Witten guy really that smart? He looks a bit phony. \_ Dunno about string theory, but string practice: ~john/ringback.jpg \_ Hot. Who's she? \_ why the fuck do people keep talking about this goddamn show? If you want to know about string theory, for some godforsaken reason, read a fucking book. \_ somepeople want a lay person's explanation to be done in an hour. Books take much longer than that. \_ How about http://staff.science.uva.nl/~rhd/string_theory.html \_ I'll sum it up in two lines on the motd: If a theory is unrelated to experiment, it's not physics, it's philosophy. for more information type "dict wank." \_ Apparently you're a wank wannabe scientist who's never read Kuhn. williamc \_ I heard there may be experiments with the potential to falsify string theory coming after the CERN accelerator comes online in 2006.  ilyas \_ right, and when they do, the theory will either be falsified or just unverified. wake me when they can calculate the mass on the electron from frist principles or predict a new particle acurately, or do *anything* predictive. <snore> \_ We will eventually run out of things to predict. A theory isn't good only if it predicts something new (although that's really nice). A theory is good if it doesn't contradict any data and is as small as possible. Personally I know next to nothing about string theory, and lack the background to learn more. I don't know how well it fits, and I don't know how small it is (or why there's so \_ How about http://staff.science.uva.nl/~rhd/string_theory.html much hype).  ilyas \_ wow, this is the first time i heard this show mentioned. i must be out of it. anyway, go read Brian Greene's "The Elegant Universe." it gets pretty dense as you get into it, but given enough dedication, you can follow what he's writing. \ the Witten/Schwartz/MGreen(not BGreene) is a pretty standard serious work on string theory: http://csua.org/u/557 witten solved a problem a bunch of other people were stuck on [i think this is descrived in vague terms in the show, but i saw it a only in part and a while ago] annd he's not doubt a bright guy ... but personally i find s. weinberg more impressive and certainly more articulate. "dreams of a final theory" is a more accessible but still interesting book. it's also cheeper than the GSW book ... which is a $50 "paperback" and fairly tough going if you dont have say 2yrs of grad math. psb \_ did anyone in this thread express interest in a "standard serious work?" \ dear mr. too short: "phony physicists rarely write standard serious works". psb \_ fuck off. real physicist \_ you lie. a real physicist wouldn't call herself such (maybe "physics grad student" or "physics prof"). i wont make judgments on whether she'd be posting to the motd. \_ doh! you got me! it turns out that i'm the pompus ass sysadmin knowitall who learns about the most useless theory in physics to impress girls at parties, and you're actually the physicist! my bad! 
2003/10/3031 [Science/Biology, Computer/Theory] UID:10878 Activity:nil 
10/30 Yahoo! News  Robots to Gain Eyes in the Back of Their Heads htt://csua.org/u/4um It reads " But as computer scientists at the University of Maryland proved mathematically in 1998, if robots could see in all directions they would not need any other sensors." What kind of mathematical proof would that be? How do you go about proving something like this mathematically? \_ why doesn't evolution favor eyes at the back of animals heads? \_ Prey animals usually have very widelyspaced eyes and can see in almost 360degrees. Predators (and humans) have forward facing eyes which give good depth perception. Why do no vertebrates have more than 2 eyes? \_ The fundamental answer guiding all evolutionary processes: efficiency. \_ only able to last 15 seconds in the sack, eh? \_ How is turning the head to look behind more efficient than procesing more signals from more eyes in the brain? \_ You save the energy needed to grow more eyes and the brain structures needed to process the extra input. \_ But you need to grow the muscles to turn the head, and for some mammals, even part of the body. \_ good point. \_ Eyes have a lot of muscles and things, at least our full functioned ones do. Also, head movement is needed anyway for eating (maybe not for humans, but for prey animals). Since prey animal eyes and hearing suffices, more eyes probably cause more problems than they help. They might also be vulnerable to injuries. Head movement is also used for smelling. \_ You might also want to take into account that more eyes also translates to more brain mass/complexity to process the information. \_ Flies have many eyes in two groups. \_ I said VERTEBRATES. \_ and mammals (and many verterbrates) have many rods/cones in two groups. \_ But flies have many separate lenses whereas mammals have only two. \_ The article explains it pretty well: "The ability to navigate was the lowest level of capability needed by a robot to work in an unknown environment, she said." "Providing a robot with "omnidirectional" vision could vastly improve its navigational skills, ..." Thus, you make the assumption that navigation is needed, then you prove that 360degree eyes are sufficient for navigation, thus it doesn't "need any other sensors." 
2003/10/24 [Computer/SW/Apps, Science/Physics, Computer/Theory] UID:10762 Activity:high 
10/24 IF AIM WAS AROUND WHEN QUANTUM THEORY STARTED http://www.makeoutclub.com/03/page_messageboard.php?topic_id=323705 hahahahahahahahahahah! maxmcc \_ Dumbass. \_ =((((((((((((((( \_ ?????? PROFIT! 
2003/10/1314 [Computer/Theory] UID:10617 Activity:high 
10/13 Second WALKING robot out of Japan. First one was Honda's ASIMO. http://www.sony.net/SonyInfo/QRIO Is anybody else even in this field? \_ learn to scroll. Specifically, downwards. \_ More importantly, why bother being in this field? Is building a robotic version of a human really all that useful? I'd rather have numerous cheaper specialized robots than 1 general purpose one that can sort of do a number of things ok. Like it would be nice if my vacuum cleaner would just turn on and clean up when I'm not around and then go back to it's closet. \_ Or a robot grammarian. \_ its god damn it. it's not hard \_ I think such vacuum cleaner aleardy exists. \_ Or a robot grammarian. \_ Wait until they have sex slave robots made in the image of your choosing. Then you'd have something. \_ Wasn't the first one a Honda/Sony joint venture? 
2003/9/34 [Computer/Theory, Computer/SW/Unix] UID:10050 Activity:nil 
9/5 I have a relatively large mailing list archive in a typical unix's mbox format. Is there any tools that allow me to view it easily, i.e. reconstruct threads, etc? \_ mutt f file (then hit "ot" to sort by threads) \_ From a mh perspective, you can incorporate it into your mail box with inc, and then do a subject field sort with 'sortm textfield subject'. ERic 
2003/8/292004/2/14 [Academia/Berkeley/Classes, Computer/Theory] UID:12244 Activity:very high 
2/13 How about a little academic antibragging? Let's poll how low your SAT/GRE score was and how well you did in school. Do not reply if you had a high score. For me, I was an EECS major and I had a SAT of 1300. My math was an abysmal 680. I think I was the only EECS guy with such a low math score. I did ok in school (some As, mostly Bs). \_ In high school, 670 Verbal, 720 Math, 3.98 GPA. Graduated Cal with a 2.74. Fucking ouch. \_ I think he wanted the other way around. I know a lot of people including myself who were good in HS but fucked up at Cal. I think it had more to do with peripheral issues than actual academics. I did well when I tried, but I had a hard time enjoying and motivating myself there. Although having Cal on my resume helped with job interviews, what with my shitty experience and having to explain shitty grades I should have gone elsewhere. \_ I failed every single math class after junior year in high school. \_ I graduated with a 2.3 GPA. \_ got in with 1200 SAT, English major, graduated with a 2.8. I'm making 6 digits as a sys adm. \_ 370 Verbal, 780 Math. Graduated with honor in L&S CS with 3.65. (Now guess my ethnicity.) \_ Asian, Eastern European, or white nerdling who lived in a cave. \_ 540 Verbal, 670 Math, and got into EECS. I thought hell would freeze over before I got into EECS with that score, but it happened. My guess is that I got in for geographic reasons. I came from Modesto, and in general the Central Valley is underrepresented in terms of % of people in California. phale \_ I scored a lot higher and didn't get in. white boy \_ Well, my SAT was okay, but my GPA wasn't so hot. I wasn't even within the top 40percentile of my class, yet I still got in. \_ are you a protected minority? \_ no. i think i simply got lucky. \_ there's a small quota for "low end white people" too \_ 640 verbal, 730 math, Entered Fall '85 EECS \_ Oh, 1370! How shameful! This is supposed to be antibragging. \_ yer supposed to say "apparently your verbal score is overinflated, because ...", or "yermom accepts all applicants regarless of SAT score" \_ 1410, l&s, noncs, below 2.7 gpa, making well into 6 figs \_ 1320, foreign student, didn't finish high school because we end in december, used 10th grade exam results to apply, our grades are like a1, a2, b1, b2, didn't know how to translate to gpa, so I just considered both a1 and a2 as 4.0, which made my gpa 4.0. they did get my transcript though so they can see everything. had one semester of 9th grade in the US and got a 4.0, so that might have helped. Cal at first wanted my 12th grade exam results, but I ask them not to be so picky, so I got in, but UCSD rejected me for unknown reasons. Cal was pretty tough for me (B), but many of my fellow countrymen did really well at Cal, most with > 3.5 gpa, one with 17 A+'s in EECS. 
2003/8/1920 [Computer/Theory] UID:29392 Activity:kinda low 
8/19 Machine learning people: what would you recommend for a reference on pattern classification? It seems Duda & Hart is the classic, but it's 30 years old  is it still relevant? Any other exceptional texts? Thanks. \_ Tom Mitchell's book Machine Learning? \_ I am told there aren't exceptional ML books. Stuart used Mitchell's book, and Bishop's neural network book. \_ if you want something not very mathematical, there are plenty of books out there that will teach you how to use a given software package to do what you want. if you want to understand what's happenning, i really like Hastie, Tibshirani and Friedman's book, the Elements of Statistical Learning. Duda&Hart is outdated and is pretty naive: there are lots of interesting links in ML, and D&H does not notice them (even the second edition, which is significantly revamped). Bishop's is very nice, but it doesn't cover a lot of topics. i haven't taken a good look at Mitchell's book, but i would pick hastie/tibsh/fried if i had to teach a class. 
9/18 
2003/8/9 [Computer/Theory, Politics] UID:29292 Activity:nil 
8/8 <DEAD>www.arghnold.com<DEAD> \_ A spanking ? Yeah.. shows their intelligence / maturity level. \_ What, did you get your sense of humor amputated at your bris? 
2003/7/3031 [Computer/Theory] UID:29182 Activity:high 
7/30 Any math major here? There was a claim that "a^n + b^n = c^n has no positive integer solutions a,b,c,n for n > 2." Ten years or so ago I heard that nobody has found yet either a proof or a counterproof for this statement. Is it still the case today? Thx. \_ Is this the theorem that some enlightenmentera math type (Fermat? Euler?) claimed to have proven in the margin of one of his journals? If so, I believe it was solved recently by a British guy using some rather advanced number theory. I'm not at all a math major, but I remember the story supposedly Fermat/Euler/Whoever couldn't have proven it because the background theory or methodology didn't exist back then, and people seemed to believe that F/E/W knew he hadn't proven it, but claimed to have done so because he thought it was provable and wanted someone else to do it. Apologies if I'm way off. \_ It was Fermat's Last Theorem, and he claimed to have a "truly remarkable proof" that would not fit within the constraints of the margin. Euler tried and failed. The British mathematician actually failed his first time, but then came back with a proof a year later that has since been accepted. It wasn't that it was impossible that Fermat had the proof, only that the British guy used something entirely different from what everybody else was trying to use, he did not use number theory. \_ It's not really accurate to say that Wiles ("the British mathematician") "failed his first time." Even in between the time a gap was found in his proof, and when he was able to fix it, most experts thought the proof was essentially correct. It would have been truly astounding if a manuscript that large and that original had been perfect the first time around. Also, I'm not sure what you mean by "he did not use number theory," since the whole paper is number theory. It is true that he used lots of concepts that were not known in Fermat's time. \_ it still cannot be proven in Star Trek TNG, so it's hopeless now. ;) \_ "I love you, but I do not love you!" \_ "This statement cannot be proven." \_ It has been proven recently, after 300 years or so of failed attempts and false proofs. Search google for "fermat's last theorem" for details. \ this was a troll, right? anyway, the people working on the FLT were really working on deeper conjectures like the ABC conjecture ... the FLT would have been a small byproduct. the people who claimed to be working on the FLT itself were probably cranks. berkeley's K. Ribet was a big player in this story. psb \_ There's an episode of Nova where he is talking about FLT at Caffe Strada. 
2003/7/24 [Computer/Theory, Politics] UID:29132 Activity:nil 
7/24 All right: http://www.emode.com/jumpto?test=eiqogt Emotional IQ test. I want to see how many of you are the wonderful human beings I know you are. 
2003/7/1921 [Computer/SW/Graphics, Computer/Theory] UID:29082 Activity:high 
7/19 http://www.discover.com/aug_03/gthere.html?article=feattech.html Maybe in another 10 years this sort of thing plus the advancement of computer graphics in general will lead to all hollywood movies being done on computer and the end of the super rich empty headed hollywood actors running around mouthing off on late night talk shows. \_ You forget that all the animated features use big celebrities for voices. Computergenerated speech will suck for a long time to come. tom \_ I'm not going to recomend it as a *good* movie exactly, but the movie "simone," with Al Paccino explores this idea. \_ Oooh! Genetic algorithms! Must ... put .. on cover ... of wired.. \_ Here Here to that. I was hoping to give them all a 9mm hemmorage 1st \_ First of all it is hear hear. Second of all wow aren't you a big man. I bet all the chicks flock to your manlyness \_ But then whose lives will we vicariously live? \_ yes we need good down to the earth "real" looking guys like Eric Raymond on the tonight show \_ WTF would anyone put a geek on a talk show? Maybe they'd start putting interesting people on them instead of entertainment figures or better yet just cancel the entire late night genre. 
2003/6/56 [Computer/Theory] UID:28643 Activity:moderate 
6/5 Someone explain the bigO notation again. Basically O(X) means a computation cannot be worse than constant*X, where X could be log, exponential, or what not. So isn't it correct to say that the bigO for quick sort could be any of the following: O(nlog n), O(n^2), O(n^n), O(n!), since quick sort will never be worse than these? \_ all of those are technically correct (except for O(n log n), quicksort is O(n^2) in the worse case). however, when your characterization of the running time is that far off from the true upper bound, that's not very useful. \_ you should look into the difference between bigO, bigtheta and bigomega notation. I believe SICP and CLR(S) both have decent introductions. twohey \_ http://www.cs.mcgill.ca/~cs251/OldCourses/1997/topic3 \_ no because that suggests it could actually be of n! time which it can never be. (unless u implement it wrong) \_ if you implement it wrong, it shouldnt be called quiksort, it shoudl be called slosort or something like that. \_ how about "spellsort" or "sloppy, lazy typo sort"? \_ If you can give a girl the bigO  youre set. \_ But only if you can do it in 0(n) time. 
2003/5/2627 [Computer/Theory] UID:28554 Activity:nil 
5/25 Dear nice math helper person, I found this from my old CLR book: Sigma k=0 to n of x^k = (k^(n+1)1) / (x1). Can someone please repost the question I posted yesterday? Thanks... \_ Almost  the "k" on the RHS should be an "x". Hint 3. Figure out how to use this to get a closed formula for 0*1 + 1*x + 2*x^2 + ... + n*x^n 
2003/5/25 [Computer/Theory] UID:28545 Activity:high 
5/24 How would one solve the following equation? Sigma i=1 to log n of (2^i * log 2^i)? I just need hints to get started. I'm not sure if I can actually use the formula n*(n+1)/2. Please help me get started. I'm not looking for an answer, just hints, THANKS! \_ Hint 1. Rewrite log 2^i as i*(log 2) Hint 2. Find a closed formula for 1 + x + x^2 + ... +x^k Hint 3. On request. \_ Ok so I tried hint 2, (2*1)+(4*2)+(8*3)+...+nlog n. Now what? \_ Try again; we want a closed formula for 1 + x + x^2 + ... + x^k, i.e, a formula without the "..." \_ uh, (x^(i+1)+1)/2? \_ Nope. \_ isn't this in first year calculus/math1a? 
2003/5/2021 [Computer/Theory] UID:28497 Activity:very high 
5/20 How important are course evaluations, and specifically the evaluations of the professors, for faculty promotions at Cal? \_ depends on the dept. all depts read them, regardless of whether the prof does. be fair and don't froth at the mouth; instructors (and depts) can ID and tend to discount wildly angry evals unless it's clear that those evals can support their claims. wildly positive evals tend to be discounted too. :) humanities gsi \_ Yes, it depends *alot* on the department. I think the Berkeley math department really doesn't give a shit about teaching. I majored in math, and got to see a bit about how that department works and how they do their hiring. When you apply for a faculty position there, they don't even ask about teaching experience. Of course there are some great teachers, like John Neu, but they decided to be great teachers on their own, with no motivation from the department. \_ And now the truth: you're an undergrad at a research university. Unless you have film of the prof molesting 6 year olds just before killing them in a satanic ritual (and his face better show) your eval won't mean anything to anyone. The translation of the above about ignoring the angry and the happy people is that the only evals remaining are the vast bulk that say very little. Think about it. \_ not ignoring all the nonbland ones, just discounting those which sound stalkerlike (at either end). there's a difference. the research univ. emphasis is correct otherwise; ucb does care more about teaching across the board than many comparable research schools, both for hiring and internally, but that's nothing to the fuzzy warmth of small colleges. one's teaching record can help for advancement but it rarely hurts enough to prevent advancement. if you've a serious contention re: someone's teaching, take it to the dept chairthey hold office hours like anyone else. \_ holy shit, no! do not go to the department chair! not if you ever plan to take a class with that department ever again. \_ If the professor is good or has tenure your evaluation forms probably aren't going to negatively affect his career much. Do you really expect one grumpy undergrad to matter that much? However a lot of professors actually DO care about teaching well and evalution forms do help them with that. Also departments who have profs that have a history of getting really bad evals tend to work with that in mind. For instance I think the famous Hurricane Wu no longer teaches undergrad classes. And that is the math getting much funcding yet expected to give almost every undergrad department which is famous at cal for not giving a damn about the undergrads. (Probably something to do with that whole not getting much funding yet expected to give almost every undergrad here a year or more of math.) Also as has been said, good evaluations help more than bad ones hurt. So give good evalutions to the profs you like, give constructive critisism to the ones you think will actually listen, and don't stress too much. \_ Mostly agreed. Bottom line, be honest, but be mature. \_ I was in Wu's Math 1B class for about 3 minutes. That's how long it took to realise the guy was a total psycho and walk out even though I needed 1B that semester. I couldn't believe no one else left. Did anyone else here actually take Wu for longer than a week? \_ Actually, a friend of mine took his upper division differential geometry course and loved it. of course he is also insane, and is now getting a phd in differential geometry. For some people, Wu is actually a very good teacher. \_ Pulled a B. Worst semester ever. scotsman \ The time has come for: http://www.jiggscasey.com/slappy/book_of_wu.html psb \_ NOW THAT IS HILARIOUS. Is Wu REALLY like that in person? All that cryptic Chinese proverb shit? \ everyone i have pointed to that page thought it was funny. everyone who had wu for a class also added "it was really like that!" The other funny thing is Wu rally cares about teaching. If you go to his WEEB site, he's written papers on *teaching* math. ok tnx. psb \_ Now you know why you have fans. \ PSB's Corollary to Godwin's Law: All conversations about teaching at Berkeley eventually end up at Hurricane Wu. psb \_ Wu, Math 113... Hard as all hell. 
2003/4/2224 [Academia/Berkeley/Classes, Computer/Theory] UID:28191 Activity:high 
4/22 Data Structure: X,Y,Z coordinates repeated. Same X,Y  different Z Ie, multiple surfaces over the same grid. Problem: Want a polynomial fit f(X,Y)=Z; save the equation in matrix/vector; then the dump into MATLAB. Question: a lot of programs do polynomial fits, but it seems to be a pain to save the equation describing f(X,Y). What can I use to do a large number of curve fittings and then save the polynomical eqn? fab@csua \_ If you took CS170 you should know the answer to your own question. Hint: it starts with an F, and runs in O(n log n) time. \_ Didn't take CS170. Care to provide an answer? The question isn't about curve fitting, it's merely about making your favorite software (Stata, SPSS, MATLAB, etc) produce a friggin' macro/list/whatever. fab \_ The data structure itself can be used to represent the fitted polynomials. You can transform between a 'set of points' representation and 'set of polynomial coefficients' representation using something called Fast Fourier Transform. I highly suggest you read up on it, any engineer should know what it is. \_ I'm not an engineer. Sorry to disappoint, but if this is routine, I wouldn't mind paying some undergrad a very modest sum to do this for me. fab@csua \_ Dear fab@csua. Are you really stupid enough to not realize you just offered someone a very modest sum to do nothing at all? The whole point of FFT is that your original matrix is a perfectly valid representation of the fitted polynomials. \_ I think it's a reasonable guess that anyone not clever enough to indent motd correctly may not be clever enough to do FFT's. \_ I took 170. We didn't cover that. Must be new math. \_ Boy, there are a lot of wrong answers here. FFT does NOT fit polinomials to data. It fits discrete complex sinusoids to it. This almost certainly isn't what the person wants. Your belligerence is unwarranted, and surpassed by your ineptitude, mr. fft guy. The op should consider performing the polyfit in MATLAB, instead of worrying about how to dump the result to MATLAB. the polyfit function in matlab does this. Also, I don't understand your question, so what nivra said ali. \_ http://www.mapleapps.com/categories/mathematics/algebra/html/FFTPolynomial.html \_ this teaches you how to use the FFT to multiply two polynomials. the trick is based on the fact that convolution of the poly coeffs is the same as multiplication of the DFT coeffs. it has nothing to do with polynomial fitting. did you just google for "fft and polynomial" and post the result on the motd? ali \_ I don't get your question. It seems you're asking for the following: You have nmesh like 2D surfaces, you'd like a polynomial fit for each surface, resulting in nsets of polynomial equations Z = f(X,Y), You'd like to save all n equations easily. Most polynomial fits should give output in terms of coefficients. These coefficients form a vector: eg. Ax^2+Bx+Cy^2+Dy+Exy+F. You will end up with n sets of coefficeints. You can save this as a big matrix in Matlab. What's so difficult about this? If you want to do multiple polynomial(2nd order  10th order) for each mesh surface, you run this whole thing 9 times, and get 1 matrix for each order, with higher order polynomial fits having many more vector elements. nivra 
2003/4/11 [Computer/Theory] UID:28084 Activity:nil 
4/11 Cringely's got an entertaining column up. http://pbs.org/cringely 
2003/3/20 [Computer/Theory] UID:27763 Activity:moderate 
3/19 Just in case some of you haven't seen this yet, there are two new attacks on RSA in OpenSSL 1. Timing attack on RSA keys: http://crypto.stanford.edu/~dabo/papers/ssltiming.pdf OpenSSL advisory is here: http://www.openssl.org/news/secadv_20030317.txt 2. Extened Bleichenbacher attack on RSA with PKCS #1 v1.5 padding: http://eprint.iacr.org/2003/052 The OpenSSL advisory will be here: http://www.openssl.org/news/secadv_20030319.txt \_ In English, if I'm a low value target do I have to worry about anything? Is this stuff theoretical or script kiddy quality? \_ the dropouts speak! \_ Read the Stanford paper. They have an exploit. They also say which configurations are vulnerable. Now set up a traffic monitor/packet sniffer on your "low value target" home DSL or cable box and fire up IRC or something. Enjoy. John \_ If you are running HTTPS on your "low value target" I would recommend applying the OpenSSL patch that fixes the first attack. I wouldn't be as concerned about the second one. In any case, if you are running OpenBSD 3.1 or 3.2 there are patches available from the usual place. 
2002/12/1920 [Computer/SW/Languages/Java, Computer/Theory] UID:26854 Activity:very high 
12/18 Math help! I'm a software engineer who forgot how to do problems like this: I have a bag with 60% black marbles and 40% white. Drawing with replacement, how many marbles do I have to draw to have a 90% chance of drawing 5 black marbles (among any number of white)? I'd appreciate a formula or explanation because I might need to change those numbers. THANKS! \_ Stat 2, kids. \_ Couldn't you figure this out using your own logical reasoning? \_ No, that is why I am asking. \_ Likely wrong: * Drawing 5 black marbles on the first try is .6**5 = .07776 You want to make the chances of 5 black marbles .9, ie, * .007776 * numDraws = .9 ==> numDraws = .9 / .07776 = 11.57 ==> rounding up, you need 12 draws. \_ When you can't find the right forumale, the easy backup plan is to write a program that will do X trials for you and find it empiracally. \_ Real answer: The answer is the smallest N, such that the definite integral of the binomial distrubution function with parameters n = N, \_ too bad you can't take a definite integral of a binomial distribution  it's discrete. \_ What the hell are you talking about? Of course you can. Just take the sum. You are an idiot. Go away and kill yourself now. \_ So do a discrete summation instead \_ Too bad there are only five marbles in the bag (3 black and 2 white) making a solution impossible. Unless you know the number of marbles in the bag, you can't give a reasonable answer. As it approaches a very high number (not infinity per se), the above solution becomes more valid. p = 0.6, taken from 5 to N >= .9. Feel free to play around with many java applets of the binomial distribution to find out what N is. \_ you can also use the chernov bound to bound this integral from above. if you're interested in these specfific parameters, you should implement what the above says. \_ It's a poly time algorithm to find N, no need to bound. \_ The question is asking drawing exactly 5 out of N, so should the answer be just: 90% = C(N,5)*0.6^5*0.4^(N5)? 
2002/12/10 [Computer/SW/Security, Computer/Theory] UID:26779 Activity:high 
12/9 Story on Blum at http://www.nytimes.com/2002/12/10/science/physical/10COMP.html Question  why is Blum "Professor Emeritus" of the CS dept when in fact he was happy enticed from Berkeley and is now ensconced at CMU with a full and productive lab? \_ Hey man, like what a traitor! I can't believe that! I'm just like ya know totally stunned and completely bummed! And like he ya know stole an emerity thingy from us! Man! \_Who said anything about him being a traitor? Just wondering why he has this title, which means "Retired but retaining an honorary title corresponding to that held immediately before retirement" when he's anything but retired. \_ From Webster's Revised Unabridged Dictionary (1913) [web1913]: Emeritus \E*mer"i*tus\, a. [L., having served out his time, p. p. of emerere, emereri, to obtain by service, serve out one's term; e out + merere, mereri, to merit, earn, serve.] Honorably discharged from the performance of public duty on account of age, infirmity, or __long and faithful services__;  said of an officer of a college or pastor of a church. \_ enticed by his wife, no less. \_ yet another reason why marriage is evil.  bdg fan #3 
2002/11/1516 [Computer/Theory] UID:26557 Activity:low 
11/15 New developments in quantum crypto: http://news.com.com/21021001965957.html \_ new? hardly. would you like a list of references from the last several years of experimental advances in quantum crypto? moron reporters report all science as a "breakthrough" to get the attention of their moron editors. \_ Even if it was new New NEW! it wouldn't matter one whit to you as the man on the street. This is only good for big governments and corps with *lots* of money. It does not work over the public net. \_ ah. but i am not the man on the street, and it does matter to me. I work in this field, which is why i get annoyed by shitty journalism. if you want to follow what's happening in the field in real time, see the links from the quantum coherence section of http://www.vjnano.org or just check Science and Nature regularly. \_ The rest of us are unlikely to ever see this in real life. 
2002/11/78 [Computer/Theory] UID:26462 Activity:kinda low 
11/7 Anyone have a good intro (i'm in math54 atm) to bayesian analysis that a college sophomore could understand? i read that the new spamassassin uses it (like ifile) but cant find a good intro text on it other than: http://www.paulgraham.com/spam.html (which was actually pretty good) \_ If you are talking about that stupid /. thing, it's really simple. It's just counting words. If a document has lots of words deemed 'spammy', it does not get through. This is literally all there is to this 'bayesian' approach. AI people call this the 'naive bayes' algorithm. \_ i dont read /. op \_ does spamassassin actually use this already? If so, how do I slap it around and tell it which is spam and which isn't? \_ the beta/experimental release (2.5) does. \_ Bayesian analysis fundamentals are simple. It all drives off of Bayes Theorem (mathworld has one, but there are plenty of better ones, if you google for it: http://mathworld.wolfram.com In a nutshell, given the observation of data D, the probability that a hypothesis is true is equal to the conditional probability that the data is true given the hypothesis * the PRIOR probability of the hypothesis being true. In essence, it ideally redistributes your prior probability distribution based on the data observed. Bayes Theorem: P(H(i)D) = [P(DH(i))*P(H(i))] (formatd was here) where  D = observ. data Sum(j=1>n,P(DH(j))*P(Hj)) H(i) = hypothesis i 1<=i<=n nivra \_ What's the relation between Bayes Theorem and Neural Networks? \_ No relationship at all. Neural networks are function approximators that work using hillclimbing. Bayes theorem is a relationship between conditional probabilities of two events (it's not actually a theorem, it just follows straight from the definitions).  cardcarrying bayesian \_ Bayes Theorem: Conditional probability. I don't remember Neural Networks using Bayes Theorem. However, Bayesian networks use Bayes theory. \_ I don't know much about Neural Networks. Presumably, they use different learning algorithms for deciding what information to weight, and how to weight its nodes/connections. Thus, presumably, you could have a Bayesian learning algorithm that decides how to reweight its connections based on observed data. nivra \_ google for bayesian networks or bayesian belief networks. Neural Networks typically have an activation function at the nodes to weight things. The Backpropagation Algorithm is typically run. I don't think I've heard of a variation of Neural Networks algorithm which uses Bayes. 
2002/11/78 [Computer/Theory] UID:26461 Activity:high 
11/7 So, when and where is this LP breakthrough coming out? \_ The paper was submitted for publication this year. I went to a talk today given by the guy who came up with it  the algorithm is beautifully simple. Unfortunately, I think he is trying to patent it. \_ Where was the talk? Where is it being submitted for publication? Who is the author (or at least institutional affiliation)? \_ At my school. The author is Andy Mirzaian, from York University in Canada (received his PhD from Princeton a while ago). \_ Did he give details about this algorithm at the talk? All the details? Were people convinced? Has this work been subjected to formal peer review? \_ He described the algorithm, but not the proof that it runs in poly time. It has not been peer reviewed yet (since it has only been submitted, not published). \_ Ah. May I ask which school you are at? \_ You may. \_ Which school are you at? \_ http://www.cs.yorku.ca/~andy/pubs/index.html google \_ Yea, got that. 
2002/11/67 [Computer/Theory] UID:26446 Activity:very high 
11/6 Strongly polynomial time algorithm found for linear programming. News at 11. \_ Uhm, that was 1979. Or are you using some nonstandard definition of "strongly"? \_ 1979 ellipsoid algorithm was weak polynomial time, i.e. it was polynomial in the size of the constraint matrix, not polynomial in the size of largest number in the matrix. The current algorithm is polynomial in both, which is the standard definition. What definition are you using? algorithm is polynomial in both, which is the standard definition. What definition are you using? \_ The same one; I hadn't considered the numerical behavior of the data. So, enlighten me, are you saying the ellipsoid algorithm can't handle in polytime, say, a matrix with floats in [0,1]? \_ No, the ellipsoid algorithm can't handle in polytime a matrix with really really large numbers. An easy example of weakly polytime problem is factoring a dot product vector of integers. If the integers are small, this problem is clearly solvable in time linear in the length of the vector. The problem is that factoring integers is hard, so there are no (known) solutions which have polynomial runtime in the size of the largest integer in the vector. The ellipsoid algorithm has a similar problem with numbers in the constraint matrix. \_ You misunderstood my question. Yes, I know what you mean by "strongly polytime." What about LP over data in a fixed range of reals? mean by "strongly polytime." What about LP when all parameters are in a fixed range of reals? \_ I am not sure. If the float is very small, or has high precision, it will take a lot of bits to represent it. It may be that in fact the ellipsoid algorithm has exponential behavior for any number that needs a lot of bits, large or small. (Note that this clearly isn't true for factoring). \ is this one of stephen smale's problems for the next century? psb 
2002/10/2224 [Computer/Theory] UID:26282 Activity:high 
10/22 http://www.csua.berkeley.edu/~ilyas/problems/coloring If you know the answer, please email me, but don't give it away here.  ilyas \_ Don't ask us to do your homework for you, mr. cheatypants! \_ I solved this already. I also received one solution from someone on soda.  ilyas \_ "Small" as in fewest vertices or "small" as in some notion the graph even need to be planar? sufficient for planar graphs.  ilyas of Euclidean diameter or something like that? Also, an edge can exist if and only if the distance is exactly 1? And can edges intersect (provided the graph is still planar)? Hell, does the graph even need to be planar? Can 2 vertices be assigned to the same position in the plane? \_ Small in a sense of fewest vertices. An edge can exist if and only if the distance is exactly 1. Edges can intersect, the graph does not need to be planar. Two distinct vertices cannot occupy the same position on the plane.  ilyas \_ what's a 4coloring? Does this question need more explaination? \_ A 4coloring is an assignment of colors to vertices, such that no more than 4 colors are used, and no two adjacent vertices share a color. The terminology I am using comes from a related problem asking whether 4 colors are always sufficient for planar graphs (which was solved by Appel and Haken with the aid of a computer in 1976).  ilyas 
2002/10/2122 [Computer/Theory] UID:26269 Activity:moderate 
10/21 By transferring 3.47 gigabytes over 3000 miles (4810 km) of network from Eugene, Oregon to Syracuse, New York in one hour, the team set an I2LSR IPv6 category record of 39.81 terabit meters per second. is their math wrong or am I dumb? from http://csua.org/u/43a \_ Their math is fine, assuming they meant, to use the ISO terms, 3.47 gibibytes and 39.81 terabits. note that the former is in units of 2^30 bytes while the latter is in units of 10^12 BITS. \ do you know what their %age of rexmits was and checksum failures? psb \_ Damn. I could stomp this record with a truck full of hard disks. \_ Never underestimate the bandwidth... \_ Netflix transfers way more data. Except they use DVD's instead of HD's and postal carriers instead of a station wagon. \_ Not in 1 hour coast to coast. 
2002/9/18 [Computer/Theory] UID:25923 Activity:moderate 
9/17 I have problems with graph theory & combinatorics cuz I didn't pay attention when I was in school. What's a good book to relearn these things (and not to get confused when looking at http://mathworld.wolfram.com) ? Mainly I just wanna understand the Ramsey Number (NP problem) alum \_ Take a look at Modern Graph Theory by Bela Bollobas, and \ hello, there is [was?] a copy of Bela Bollobas: Graph Theory: An Introductory Course [n.b. this "introduction" is a Springer GTM] at the 9th ave Green Apple Book Facility for <$10 ... It's probably worth a trip to save $30 ... I own one but probably should have bought it to "arbitrage"; If the book is not useful to you, you can almost certainly make money on it. Is that the book you mean above? This is a useful reference for some of my work but I dont think it is really a "guide for the perplexed". psb \_ Modern Graph Theory is a revised version of this, with more than twice as many pages. \_ that's twice as much learnin'!! A Course in Combinatorics by Van Lint and Wilson. (The latter doesn't have much on Ramsey numbers though.) ok \_ CLR? \_ I've seen some copies of this pretty cheap on http://half.com. 
2002/9/15 [Computer/Networking, Computer/Theory] UID:25898 Activity:nil 
9/15 I want to block http://advertising.com, http://doubleclick.com, http://fastclick.com, etc etc etc on my router. Is there a blacklist I can find? Thanks. 
2002/9/910 [Computer/Theory, Computer/SW/Unix] UID:25822 Activity:high 
9/9 Let's say I am writing a multithreaded email client and I want to allow for messages to be sorted. Problem: to sort elements I need all elements and if this thing is supposed to scale retrieving all messages when changing a sort order will not work. How do real world mail clients or applications in general deal with this? \_ you only fetch the headers and sort based on that info... thats how current clients do it.. you fetch the body when you actually want to read the message. shac \_ I think he was asking "how do i sort this if I've got incoming email at the same time I'm sorting?". \_ i actually read that as "retrieving all messages every time i change a sort order..." which makes more sense... all you need to do is before you resort check if there are new message headers, if so then add them, THEN resort. some mail clients dont even do that.. you hafta tell it to check for new messages, then resort by date or subject. shac \_ yup, that's it. The problem is not resorting but (avoiding) retrieving all messages because there can be a lot of them and if there are many users there goes the neighborhood. I can't retrieve all messages because it's a webbased client with pagination (think OWA) so there's a space vs time tradeoff and I am always fetching headers a page at a time. IMAP serverside sort should solve this but apparently it's still in experimental stages... OP can be a lot of them and if there are many users there goes the neighborhood. I can't retrieve all messages because it's a webbased client with pagination (think OWA) so there's a space vs time tradeoff and I am always fetching headers a page at a time. IMAP serverside sort should solve this but apparently it's still in experimental stages... OP [ reformatted  motdformatd ] \_ Your bad English grammar fu is superior to mine. I bow to your superior ability to decipher nonEnglish texts and extract meaning. \_ no idea how commercial apps do it but as a first cut how about you sort what you've got and then do an insertion sort on any new mail? Most folks are mostly sorting on date received so you'll just end up sticking new mail at the front or back. If you know what you're sorting on such as date you can take advantage of that fact to speed things up and "cheat". 
2002/7/2628 [Computer/Rants, Computer/Theory, Computer/SW/Unix] UID:25428 Activity:moderate 
7/26 My computer monitor stopped working recently (screen is completely snowy white). I want to get rid of it, but I'm not sure if throwing it in the garbage is the best thing to do. Is there any sort of place that takes in nonworking computer equipment like this? I'm not in the Bay Area. \_ try throwing it off the top of a parking garage. make sure you have two escape routes or more. \_ eBay? \_ Depends on your community. Your dead monitor has lots of lead in it. See if your local garbage collector has a program that deals with dead TVs and monitors. \_ your monitor is officially hazardous waste, to just toss it in the trash would be breaking the (poorly enforced) law. \_ Look around. There are dead computer collection events every 2 or 3 months in my community. I heard some of these get shippied to China and dumped in inappropriate places, but hey, that's not our problem right? \_ computer recycling center: http://www.crc.org \_ I wish I knew this before I threw away my 386 several years ago. \_ I also have a broken monitor...it's missing the color blue... does anyone know the best way to fix this? Thanks. \_ Did you accidentally mess up the color balance setting on the monitor? It happened to my cousin once and that took me a while to figure out. \_ Ship it directly to China. That's where all the warm fuzzy feel good "recycling" places are taking theirs. \_ Throw it in da Bay. \_ toss it in your neighbor's dumpster. \_ oooh, good call. I wish I'd thought of that first. 
2002/6/2728 [Computer/Theory] UID:25227 Activity:high 
6/27 Can anyone who has used Maxyma and Mathematica/Maple comment on them? I am using mathematica and find it overhyped yet inconvenient for the kind of symbolic manipulation I need. I need to implement diverse kinds of say algebraic strctures and rules and to exert close control over how the the rules are applied. That is I know how best to simplify but want to use the program to do the bookkeeping. Ok tnx. \ er do you actually have a copy of MACSYMA ... I'll buy you lunch if if you actually have VAXIMA running on a VAX. Mathematica used to be really unreliable. I havent followed discussion about it for a few years and mostly use it do draw pictures or do pretty simple things now [some factoring and computing some sums] ... I like Maple6 and many colelagues use matlab. psb \_ Derivatives of DOE Macsyma is now GPLed, see e.g. http://www.ma.utexas.edu/maxima.html and http://sourceforge.net/projects/maxima and on GNU. No I haven't tried to install. I want to have some confidence that it will be useful before trying that. \_ for symbolic math, Maple is great. matlab is only useful if you have the symbolic math toolbox. \_ what do you all mean by "symbolic math"? symbolic as opposed to what? does this mean like modern/abstract algebra? \_ for example: symbolic: finding the indefinite integral (a symbolic expression from another symbolic expression) nonsymbolic: finding the value of a definite integral (numbercrunching) \_ Anyone know of a good open source version of a program like matlab or maple? 
2002/5/2425 [Computer/Theory] UID:24928 Activity:moderate 
5/23 For those of you who missed the "commanding heights" on PBS, you can actually watch the whole series online. http://www.pbs.org/wgbh/commandingheights I hope that all PBS shows are online in the near future. \_ The url for the qt movies is: http://www.pbs.org/media/commandingheights Its pretty cool that PBS put the show on the web. I'd like to see other PBS series on the web as well, starting with American Experience and Nova and perhasp the recent Evolution. \_ From the "STORYLINE" link: "For more than half a century the battle of ideas will rage. From the totalitarian systems to the fascist states, ......" What's the difference between totalitarianism and facism? \_ Fascism is a form of totalitarianism with some additional characteristics. \_ cool... do we really still have any libertarians left? \_ We never had any real libertarians. 
2002/5/810 [Computer/Theory] UID:24763 Activity:nil 
5/8 Which class teaches you polynomial arithmetic? \_ Math 32, remedial precalculus. Or did you mean the study of formal polynomials? Then it's Math 113, some incarnations of 114, probably parts of 250A, and 251. alexf 
2002/5/89 [Computer/Theory] UID:24748 Activity:high 
5/7 Had this argument last week with a snobby coworker of mine. He argued that the diff algorithm isn't linear. Why isn't it linear? \_ what a lame argument. \_ try reading it. /usr/src/contrib/diff \_ punch him in the nose and point and laugh as blood runs down his face. \_ This is the algorithm used (according to diff's own docs): http://www.cs.arizona.edu/people/gene/PAPERS/diff.ps As you can see from the abstract, the worstcase performance is linear in the size of the file, but also dependent on the size of the minimal diff, D, with O(ND) being the actual bound. Whether that bound can be referred to as "linear" is a fuzzy question. Note also that the averagecase performance in a supposedly reasonable model is O(N+D^2), which is "linear in N" by most people's definition of the term. alexf \_ Fuck that. Punch him in the nose instead. 
2002/5/78 [Computer/Theory] UID:24746 Activity:high 
5/7 Say you have a matrix A, and they say that A * A^1 = 1 (right?) How does one get A^1? \_ you'd think if you were going to post some simple math question, you would first read the replys to the previous one. http://www.mathworld.com has the answeres to all of these questions, with links to related questions, great examples, etc. that should be your fist stop, not the motd. \_ one takes math 54 \_ obviously A^1 is the same a 1/A. \_ for a 2x2 matrix, it's trivial. for larger matrices, check out a linear algebra reference for details. but for now: (a b) ^1 = M(d b) where M = 1/(determinant) (c d) (c a) 
2002/5/46 [Computer/Theory, Computer/SW/Security] UID:24704 Activity:high 
5/3 If I want to learn about error correction, compression, and cryto, which class would I take? crypto? _/ \_ Info theory at Stanford. Berkeley does not teach ugrad info theory. \_ Information theory. Read Thomas & Cover. There is an information theory class using that book at Stanford. Berkeley does not teach information theory to undergrads. \_ 170 talks about the basics of both, 150 has some error correction too. specifics? \_ Crypto classes: 261 (well, security), 276 (protocollevel), and this semester Wagner taught a 294 which was blockcypher level. Even though I've managed not to pay any attention to 174, I remember somebody saying something about entropy, so likely has to do something with compression and/or random number generation. chialea \_ "managed not to pay any attention to 174". Okey dokey, now who was making noise before about the best Cal ugrads not getting into Cal grad school? \_ Not best. Schmooziest. Big difference. \_ if you were as good as chialea, wouldn't you be bored by 174? chialea #1 fan \_ several EE courses discuss compression (the multimedial related signal/image processing courses) \_ depending on the prof, Math 114 often covers coding theory, and errorcorrecting codes.  rory 
2002/4/28 [Computer/Theory] UID:24619 Activity:very high 
4/26 Is it possible to find the median value of a group of numbers in O(N)? \_ Yes, but in practice people use an O(n log n) solution. op \_ Yes; there is a fairly simple randomized algorithm which does this with high probability, and a nontrivial deterministic algorithm that people don't usually bother with. If N is really large, it's probably worth your time to look into the former (see section 3.3 of Motwani&Raghavan, where it's called "LazySelect"; the original paper is Floyd&Rivest, CACM 18:165172, 1975, but it may well be rather unreadable; M&R also gives refs for the deterministic algorithms). alexf \_ you can sort numbers in O(n) with radix sort and counting sort. \_ sorting is normally assumed to be O(n lg n) \_ and your point is...? 
2002/4/2324 [Computer/SW/Database, Computer/Theory] UID:24554 Activity:high 
4/34 Most databases use Btree structure to store the data. Is there a mathematical proof that this is the most efficient way to store a dbase? Best in update times, delete times, record sizes, etc? I'm wondering if there are better structures for specific structures. Like a database full of numbers is different from a database of words. thanks. \_ Proving lower bounds on storage space of an arbitrary object is undecidable (it reduces to finding the Kolmogorov complexity of the said object).  ilyas \_ You meant "reduces from," of course... (the other direction does not imply uncomputability; and there are other issues with storage of arbitary data structures [e.g. efficiency of later access]). alexf \_ Sorry, I wasn't using 'reduces' in the technical sense, I just meant the two problems are equivalent, but yes. If I had a terminating algorithm for proving lower bounds on storage space I could use that algorithm to find the Kolmogorov complexity of an object, which isn't possible. Interestingly enough, some lower bounds on running time of data access exist (O(log n)).  ilyas \_ Uh, the insert/delete/search bounds are well known and well understood. STFW or read a databases book to find out and draw your own conclusions. They are not by any means the best, but, they are good for many of the more common operations that are done on a database, and not too bad (TM) on the less common operations. And the fact that you ask if the performance would be different for a database of numbers vs. words shows a gross misunderstanding of what the hell a database is. 
2002/4/18 [Uncategorized/Profanity, Computer/Theory, Computer/SW/Apps/Media] UID:24474 Activity:nil 
4/17 for a while various heavy metal bands were hiding satanic shit in their songs that can only be heard if you play the record backwards. does anyone know of a metal song that displays some satanic looking shit if you take the fourier transform of the song? that would kick ass. i'm thinking of emailing Danzig to suggest it. \_ <DEAD>chaos.yerbox.org/face<DEAD> and http://www.tp.spt.fi/~cleth/aphex.gif \_ that kicks ass!@!!! thanks! now much time will be wasted combing my metal mp3s. 
2002/4/11 [Computer/Theory] UID:24414 Activity:high 
4/10 What is the class to take to understand inner/outer product, vectors, norm induced, orthogonality, etc? \_ What you need to do is use your drug of choice while studying and then use the same drug shortly before the exam so you'll be in the the same state of mind and you'll be just fine. \_ It was Math 54 in my day. (That day being the one after Math 50AB.) geordan \_ Math 50AB is gone... but 54 is still the one this guy's looking for. \_ Given that the op said "understand", the answer you probably are looking for is 110. If you meant "understand well", the 250AB series (B in particular) would probably be useful as well. Or at least H110. 54, at least as taught by most profs these days, is too focused on teaching basic computational skills to give people a good feel for the concepts. alexf \_ i did not take 54 at cal, but my observation was that it was taught in some way that just failed to work at all. i asked reasonably smart people if they know what an eigenvec tor was after that class, and they had no clue. if you want a lower division introduction to linear algebra or vector calculus, go to a JC where they hire actual teachers instead of mathematicians who are forced to teach. \_ Sorry but I learned what an eigenvector is the first time Kahan explained it in class. What's wrong with you? \_ Geeze. Eigenvector: x such that Ax = kx, where k is the eigenvalue.... \_ What does it *really* mean, though? That was explained to me first in Math 50B, although I already knew how to compute eigenvalues from Math 50A (and even high school). So, as always, YMMV. dim \_ What an eigenvector *really* means depends on the matrix it comes from. I've taken both math 54 and math 110, and true, they don't teach intuition. But then again intuition is not something that can be taught, but something acquired through application. For a good introduction, which seems to be what the poster is asking for, read "Linear Algebra and its Applications" by Strang. If you want to see linear algebra in action, take a computer vision course (CS 280) or a course on semidefinite programming/convex optimization (EE 227 I think). Through all these courses, the important thing to keep in mind is that linear algebra is a *framework*. It's a compact way of representing a certain kind of problems. Because this model is welldesigned, there are certain properties that has some meaning in real life. Eigenvectors are an example.  alice 
2002/4/10 [Computer/Theory] UID:24397 Activity:moderate 
4/9 i am working on modeling semiconductor devices and i need a curvefitting tool to generate mathematical model of the devices. what are some of the free and commercial tools available? thanks. \_ Get a book, copy the algorithms. \_ excel works quite good. thx anyways. op \_ You can do curve fitting with neural nets  a couple of lines of C, and even fewer lines of matlab. \_ just a couple lines, are neural nets that simple? (don't know anything about them). \_ sure, provided you have the right library. \_ Libraries are for the weak. 
2002/4/56 [Computer/Theory] UID:24340 Activity:low 
4/4 Does anyone still have the url of the webpage about the infamous Professor Wu of the math department? Thanks! \ /tmp/hurricane_wu psb \_ http://www.google.com/search?hl=en&q=book+%2Bof+wu \_ Thanks guys! You the best! \_ oh the memories... I still can't believe I almost got a B scotman scotsman 
2002/3/314/1 [Computer/Theory] UID:24276 Activity:very high 
04/730 I forgot my calculus, please help me out. does the line integral of a function f(x,y) compute the length of the function on top of that line, or the area of the function above that line? (e.g. does the line integral of a hemisphere calculate the length of the arc above the line, or the area of the halfcircle above the line?) \_ i predict that you will be much more satisfied with the responses to your question of you tell us *exactly* what the problem is. \_ I did already, look in the parens. \_ the "line integral of a hemisphere" makes no sense. one takes the line integral of a *function* along a *path*. \_ a hemisphere is a function. \_ a hemisphere is not a path. a halfcircle is perhaps what you're thinking of? \_ You want a surface integral if you're integrating over a surface. \_ asking these type of questions on a sysadm infested forum will get you nowhere \_ i am not a sysadmin, and i have a math degree. it's justa badly phrased question. \_ ditto. The MOTD: Not Just For Taos Monkeys Any More (tm) \_ Isn't Taos dead? \_ Taos Never Ended 
2002/3/19 [Computer/Theory] UID:24153 Activity:kinda low 
3/18 Looking for light reading material regarding quantum computing, thx \_ Try Feynman's Lectures on Computation. \_ Anything by Rand. 
2002/3/18 [Computer/Theory] UID:24144 Activity:very high 
3/17 What are some good schools that do quantum computing research? \_ Stanford, Berkeley, MIT. \_ Stanford only has Colin Williams and not much else. Berkeley and MIT are probably the main hubs right now. There's also Andrew Yao at Princeton, some people at Cambridge and Weissman IIRC, and the rest are at Almaden and some of the national labs. Why do you ask? Is there some particular flavor of quantum you're interested in? alexf \_ interested in the abstraction and application layer (eg simulator based on primitives defined by physicists, etc) Got comments? \_ Ehhh... Based on my rather limited understanding of the field, it will be a decade or two before the application and abstraction layer become relevant. The application and abstraction layers become relevant. The highestlevel work that exists right now is using physical primitives to get algorithmic primitives, and even then the algorithmic primitives under consideration are rather abstract and rarely of any practical value. Also, quite a few people are working on lower bounds, at an equally abstract level. Perhaps you may find the following tutorial on quantum computing interesting: http://xxx.lanl.gov/pdf/quantph/9809016 \_ I was unable to view this URL. Maybe another, or perhaps I'm misunderstanding a cue in how you wrote it? While written a few years ago, it is still showing the state of the art insofar as the practical aspects of quantum go. alexf \_ Shore's at AT&T Labs, too, but no one else.  chialea \_ Lea, it's Shor, and there're other quantum people there, including a few fairly wellknown names (Sloane, Rains). If I were you, I wouldn't make blanket statements such as the above on a topic I haven't worked on much...alexf \_ Aww, cut the bitch some slack  she hasn't finished E190 yet, after all. \_ Yet another anonymous behindthecomputerscreen tough guy. \_ why not? you do. :P chialea \_ Mmm? Explain? alexf \_ wow look at this great online rivalry. \_ Seen better. Both are off the cuff, one is arrogant off the charts, the other is motd normal. The arrogance holierthan thou thing has really been done. Very boring. 
2002/2/8 [Computer/Theory] UID:23817 Activity:nil 
2/8 tonight on PBS: http://www.dms100.org/worksucks 
2002/1/2122 [Computer/Theory] UID:23618 Activity:very high 60%like:23621 
1/21 What's the word on the various profs teaching math 110 this semester? They are Kahan, Aschenbrenner, Pachter, Kantorovitz, Iliev, and Wogoner. Obviously taking from Kahan is somewhere between suicide and insanity (probably a little of each). What about the others? \_ If someone ever tells you, "Prof. XYZ is really hard but you'll learn a lot!" you should run, not walk to the nearest L&S advisor to drop the class. Don't ever take a class with a Prof described that way. \ taken a class from Marina "you'll probably fail but she's one \_ i thought all math profs were gay? of the best teachers i had" Ratner? \_ she told me "i think probably you will not get an A, but don't worry." after i failed the first and second midterm. she gave me an F. I rally am glad i took it though, since i learned lots of set theory that i never would have learned from the guy with the weird sore on his head. \_ And now you have an F on your record and need to retake the class. Good plan, not worrying and all that. \_ WAgoner sucks hard. \_ Wu sucks so bad, how come the math department never does anything about it? Like firing his sorry ass. \_ He's a minority. Can't do that. Look at the guy who was giving secrets to the Chinese. Nothing happened because he was screaming racism and the Clintonistas broke. \_ The guy who everyone accused of being a spy wasn't even Chinese. Last I checked, Taiwan was a proUS country. I'm not crying foul because it's racist. I'm crying foul because, as usual on the motd, you accuse people without actual proof but give the bullshit excuse that "revealing evidence would comprimise national security". Trial by McCarthyism died decades ago. \_ Say what? Since when does being from a province founded by the losers of a civil war make you not of that race? That's just fucking bizarre. \_ I am Chinese and the math dept has my full support if they want to fire his sorry ass. \_ Chinese studying mathematic is not a minority group for admission or faculty recruiting. The math dept cares only for researches. If Wu got published and recognized for research, he would not be fired. You need to research more on how academic positions are filled. \_ Why? \_ Pachter is cool in person, but is known for being somewhat tougher than normal in upperdivs. Also generally considered a decent teacher, I'm told. alexf \_ Wow... I thought it was just me. Although Kahan is no Wu... \_ negative vote: Kahan: ... Wu: ...... Lam: .. Ogus: .............................. \_ I think fewer have tasted of the wrath of Kahan \_ I survived the splintery broom handles of each. Kahan was slightly less painful during the actual rape, but left me much more bloody. Wu was an experience I would do best to leave to the imagination. scotsman \ ObWu:http://www.jiggscasey.com/slappy/book_of_wu.html BTW, it's kind of pecular that Wu takes teaching so seriously. If you go to his homepage there are links to papers he wrote on teaching math, ok tnx psb \_ this makes me curious. What are some of the "inappropriate" exam questions he has raised \_ I took Wu for math 104. He was as fair as anyone else, never asked a trick question or a question that wasn't unlike anything we hadn't seen in the homework. THe fact is, he always rags on the math 1a and 1b students for not trying to truly learn math, but just trying to get a good score. He held extra unpaid tutoring hours once a week after class, and cared very much about each student putting in the effort to learn. He's not pretentious nor holier than thou with his smarts, whereas Kahan is disliked universally across the board for his rather arrogant behavior, this includes faculty. The exam difference is that Kahan will give you questions that distinguish the top 1% in a class whereas Wu (for upper div math) \ i dont think kahan is arrogant. i think there are cases where he probably thinks "i have been thinking about this for 5years so no i dont really want your off the cuff opinion." however i image many of you working in technical fields have been in similar situations. in my experience he does not say things without researching them and when he calls people wrong he provides examples and doesnt just say "i am kahan. you are wrong." i think kahan is probably the most discussed prof in themotd, which is kind of interesting. psb \_ He does have a rather selective memory though. Have a 'discussion' about something with him, and prove your point. Watch as he forgets this ever took place and have the same 'discussion' over and over with him every week. \ he probably marked the conversation to be garbage collected! will give questions that distinguish who has been following all the homework and are comfortable with the material. Also, everyone in Wu's class pretty much thought he was awesome and brought the material to life. \_ maybe he's a better updder div professor than lower div i doubt you'll find one math1a or 1b student who would agree that Wu was 'totally awesome' \_ I think too many of Berkeley professors cater to the top 10% of students rather than the other 90%. Kahan is just an extreme case of a general problem. He probably should just teach graduate level classes. \_ good way of looking at it. BTW, there are 15 seats left in the class that is truly Kahan's baby, Math 128B, i.e. "I am the most recognized badass in Numerical Analysis today, and you're all fucked." Anyone want to take it with me? \_ Heh, I know profs who call Kahan funny names behind his back \ kahan was considered kind of a oddball in the dept for a long time but he's also one of those profs who other profs in candid moments will say "you know he's really smart guy" ... i remember BH and PNH both saying that about him. The rep for being peculiar comes from stuff like when the faculty was soliciting ideas for "what would you do with a terabyte of storage" [i think this was for the sequoia 2000 proposal] and most profs had various responses saying they could use a few hundred gigs for this or that. Then they asked Kahan and he says something like "i need 1.3453245 TB to compute <>"[i dont remember what it was _/ 103 1/21 What's the word on the various profs teaching math 110 this se 19 1/21 What's the difference between a resume and a CV? 5 1/22 http://Linuxrouter.com,org,net are not working. I need a crash course on 5 1/22 Hello, here is an interesting problem [harder than the canoebowl 2 1/22 A moment of silence for Peggy Lee who died from a heart attack in he wanted to do but apparently he had worked it out very carfully and everyone else was kind of speechless]. i think there was a almost collective "holy shit" when he won the Turing Award. psb \_ imagine kahan teaching in grade school. \_ fine they're smart people. But smart != good teacher 
2002/1/89 [Academia/GradSchool, Computer/Theory] UID:23497 Activity:very high 
1/8 I am interested in handwriting recognition. In particular, Chinese handwriting recognition. What are the relevant CS/EE fields? What is a good grad school to go to for such things?  person who finally found out what he is interested in, but don't have much idea on how to pursue his interest. \_ AI. Berkeley, MIT, Stanford, CMU, UCLA, UIUC, UW, etc. \_ Machine Learning and AI.  CMU, Cornell, Berkeley, MIT, ... Here's a link to code which recognizes faces (which is similar to recognizing handwriting): http://www2.cs.cmu.edu/afs/cs.cmu.edu/user/mitchell/ftp/faces.html \_ Thanks, what about some tier two schools? Are they worth going? Or should I just find a job at the companies that are implementing these things. \_ Gradschool or not depends on where your interest lies. If you want to work on developing algorithms on handwriting recognition, then you should go to grad school (in machine learning). If you're more interested in the engineering and implementation aspect of the product, then work for a company. Somewhere in between > work as a developer at a research lab. If you decide to go to grad school, then first find a research AREA you're interested in. Going in with a specific application in mind is not really good, because 1) your interests will change, and 2) you may not find a professor who's interested in EXACTLY the same application. I would suggest you start by reading some papers on handwriting recognition, and look at the homepages of people who do ML, and pick your favorite people. Then write those people a nice email and demonstrate you know what you're talking about. If you still have a couple of years to kill at Berkeley, then start by doing some research in the area. I don't know why I'm even writing all this down. You're probably not even going to read this far.  alice \_ Thanks! That helps! You are a very nice person. I think I am interested in the engineering and implementation aspect. Does that rule out grad school or can it still be useful? \_ In that case, check out http://www.synaptics.com Apparently they make (made) chinese handwriting recognition products, according to John Platt's homepage. http://research.microsoft.com/~jplatt John, perhaps known in this crowd as the inventor of ClearType, is smart and a nice guy. He may even answer general inquiries from random people about about Synaptics and handwriting recognition. And please, don't call me nice. It gives people the wrong impression.  alice implementation aspect. Does that rule out grad school or can it still be useful? \_ If you have further questions, be VERY nice and send her an email. She might respond. Sending personal email reduces the chance that you're just trolling or interested in the hypothetical \_ why not a Master's where you concentrate on learning AI, and machine learning ideas, and maybe work on a project on machine learning, then go to industry afterwards? Another option is to take classes at Stanford's SITN (if they have those courses there) while working. \_ Study Dvorak instead. \_ Japanese Handwriting recognition demo: http://ai.bpa.arizona.edu/go/mlir/japanesehandwriting.html 
2001/12/20 [Computer/Theory, Computer/SW/Unix] UID:23318 Activity:very high 
12/19 Why is EULERPATH not as important as HAMPATH? (seriously, what is the main diff between "edges only once" and "nodes only once") \_ Euler path is a linear problem O(V+E). Hamiltonian circuit is NPComplete, which means that it's at least as hard as thousands of other problems. Produce a polynomial time algorithm for Hamilton and you can crack nearly any publickey cryptography system. \_ isn't prime factorization NPintermediate? \_ So? All that means that you don't _have_ to solve Hampath in polytime in order to solve factoring. If you _do_ solve Hampath in polytime, this still gives you polytime factoring. \_ what's the O(n) alg for EULERPATH then? \_ Find all nodes with odd degrees. If there's more than 2, halt and return "impossible". Else, start at a node of odd degree if there is one, and just keep going around while there're edges from where you are which you haven't used yet. It's fairly easy to prove that this will always find an Euler path if one exists. alexf \_ Not true. Consider this graph: _ _ /_  If you start at the top node, go down, go right, then go diagonally down and left you've cut yourself off from the square at the right. But you can start a new cycle that picks up that square, then splice it into the other path. ok \_ I stand duely corrected. Apply the algorithm listed above repeatedly to the graph, restarting at any point which has unused edges coming out. You'll get a collection of paths, of which all but at most 1 will be cycles; you can then "merge" the cycles into the path by just taking a "detour" around the cycle from the point where the path touches said cycle. This should still be linear. alexf \_ Duly. 
2001/12/5 [Computer/Theory] UID:23148 Activity:low 
12/4 If a DFA of M accepts all languages in L, then what exactly is the "complement of L"? Isn't the complement of L everything that M does not accept, which is an infinite set? \_ Usually you phrase things as "a DFA M accepts the language L" with the extraneous "in" and "of". The complement of L is usually everything not in L which is what M does not accept which can be an infinite set. This is not a problem. Consider the language L = {} (e.g. L contanis only the empty string). In this case L complement is all nonempty strings. It is trivial to build a DFA to recognize L complement. emin \_ ok thx emin. I can prove that if a DFA M can accepts all L in polynomial time, then ~L can also be accepted in polynomial time. However, how do I prove/disprove that if a DFA M can accept context free grammar, then ~M can accept all ~L? \_ DFA's can't accept CFG's. Did you mean "PDA M"? Take your method of converting M into ~M and prove that it accepts the inverse language. \_ Actually DFA's can accept any CFG which is also a regular language. For example, the language L = {} is both regular and context free. However, as you point out, there exist CFG's which DFAs can not recognize. To the original poster: I'm not quite sure what question you are asking, but the following information might be useful to you. CFGs are not closed under complement. That is if L is a CFG then ~L need not be a CFG. emin \_ YOU KICK ASS original poster taking CS GRE 
2001/11/2829 [Computer/SW/Languages/C_Cplusplus, Computer/Theory] UID:23132 Activity:high 
11/28 http://unicoi.kennesaw.edu/~jbamford/greguide/gresample.html Question 8, why isn't the answer C? 1/2 of a binary search (on average) would be (log2 N)  1, where N is the number of nodes in a complete binary tree right? So C SHOULD be the answer. But for some reason it is not. Help. \_ you should send email to help@csua.berkeley.edu. maybe they \_ you should send email to root@cory.eecs.berkeley.edu. maybe they can help you there. \_ please don't waste our time kevinm 
2001/11/2728 [Computer/Theory] UID:23122 Activity:nil 
11/27 More on letters of rec: So it was generally agreed that prof >> TA, but what if you take a class that is taught by someone who is basically a postdoc? I guess that's just somewhere in between? Also, how valuable is it to get letter from a nontech (CS, Math, etc) prof if you're applying to tech grad program? \_ Between a postdoc who knows you well and can write highly of you and a prof who doesn't care, pick the postdoc. Prestige is important, but it will be meaningless if he/she cannot write highly of you. The important thing, imo, is that the letter makes the admission committee able to appreciate your achievement. A letter from Patterson saying "he took my computer architecture course and he graduated" is pretty much meaningless. If you already have two excellent recommendations on your tech "achievement", I'll go for a humanity/math letter. 
2001/11/2728 [Computer/Theory] UID:23121 Activity:high 
11/27 Let's talk about ilyas' blueeyed problem. Again. I still don't see how the base case works. Assume 1 blue eyed person only. How would that person know that he should commit suicide? Wouldn't the brown eyed person think the same way and commit suicide? Think induction. Tom's basis is flawed. \_ The big problem with this whole thing is that it assumes that all these people have taken math 55 and understood it. \_ Please report to the food vats immediately. \_ No. The key is that there exists at least one blue eyed person. If only 1 blue eyed person exists, he will notice that as far as he can tell, no one has blue eyes. Since at least one person must have blue eyes, it must be him/her. So (s)he kills himself. \_ Added to that, a brown eyed person would see that there is a person with blue eyes, and would expect him/her to commit suicide after the first day. And when (s)he does, the brown eyed person concludes that that person was the only blue eyed person in the town. \_ Can you explain the case of three blueeyed people, both for blue and browneyed people. Won't everyone expect someone else to commit suicide? \_ Another way to explain it: [someone else motdmashed the first explanation] For two blueeyed ppl, a blueeyed person would see one blueeyed person on the first day. At the end of the second day, both blueeyed people would kill themselves. For three blueeyed ppl, a blueeyed person would see two blueeyed people on the first day. He would see them again on the second day. Assuming that there are only two blueeyed people total, they would kill themselves at the end of the second day. But we are saying that our given person is blueeyed, and there are three blueeyed ppl. On the third day, this blueeyed person would STILL see two blueeyed people. At the end of the third day, all three blueeyed people conclude that there must be three blueeyed people, and they must be one of them, and kill themselves. \_ Ah, ok, I finally get it. Thanx! stupid Math 55 flunkie For two blues and one brown, from a blue's point of view there is one other blue and one brown. Now at the second day, each blue knows he's a blue, because if he was brown the other blue would have known he was the only blue since there is at least one blue and the other two are brown. So both blues kill themselves. From the brown's point of view, he knows he's a brown by the third day, since the blues wouldn't have killed themselves yet had he been blue. 0x48 \_ Take math1a or 55 or high school math. learn induction. 0xAFB \_ one of the many problems is with ilyas' question. He should add that each person can see every other person ONCE PER DAY. This question really reflects ilyas' intelligence. \_ 1.) The kingdom was small: any person in the country could meet all others in one day. ? \_ it says 'could', but that doesn't mean it is necessarily so. \_ Rule 1 suggests that they could if they wanted. Rule 2 suggests that everyone really wants to. The weakness perhaps is we have a lot of inferences. \_ The weakness is that it's a stupid problem. tom \_ I thought the puzzle was a good one and I learned some things. Of course, the problem did have problems. 
2001/11/2427 [Academia/Berkeley/Classes, Computer/Theory] UID:23094 Activity:insanely high 
11/23 Can someone please tell me what textbook and/or author we used for: math50a, math55, and stat134? I need it from 19931995. THANKS! \_ let me guess, ypu're applying to MIT, right? \_ let me guess, you're also applying to MIT, right? Or are you already there? Help me out dude... \_ I'm applying. And I'm a dudette, btw. chialea \_ where else are you applying to? Let's get together and start a grad school application club!!! \_ chialea, what'd you get on your subject & general? \_ stfu ilyas \_ It wasn't me, moron. I already know what she got.  ilyas \_ bad things. especially the subject. and my GPA. chialea \_ 3.5 isn't bad considering the average GPA@Cal=2.3 you're off by a full halfpoint _/ ugh, you mean 2.8? 1.8? _/ it's around 2.8 _/ \_ Is that the average systemwide? how about for CS? \_ _Discrete Mathematics and Its Applications_, Kenneth H. Rosen (Math 55), don't know the rest. geordan \_ Math 50a is now Math 53, right? If so, Stewart's "Calculus: Early Transcendentals" was probably used. \_ no, 50a became 54. for math 50a in spring 1994, we used anton, _elementary linear algebra_ and boyce & diprima, _elementary differential equations and boundary value problems_. lila \_ THANK YOU SO MUCH LILA, you've been very helpful. By the way, are these classes required for the "other" CS? \_ no, just math1b and cs61a. i took math50a because i was briefly under the delusion that i wanted to study physics, and i took cs61b just for fun. lila \_ well? Did you really have a lot of fun? \_ We used the same two books (different editions?) in Spring '97. 
2001/11/8 [Computer/Theory] UID:22982 Activity:nil 
11/8 Quick math 55 poll please: Took 55, find it useful as a programmer :. Took 55, find it useful as a theorist :. Took 55, find it *completely* useless :.. Didn't take 55 : 
2001/11/8 [Computer/Theory] UID:22976 Activity:moderate 
11/7 So, okay, I'm 1/2 way through math 55 and I don't see how relevant it is to computer science. any comment? clueless sophomore \_ Take CS170 (and 172/174 if you have the time). You'll get a good whiff of why 55 is vital to CS. Even 150/164/188 will rely on snippets of it. alexf \_ Unless, of course, you take the class from Kahan. In which case, none of the class will be relevant... to anything. Unless, of course, you are a math prodigy. Then it will be relevant to increasing your nerddom. \_ Kahan will be teaching math 128B next semester, Im preparing NOW FYI, anyone else taking it, be prepared to have backup classes so you can drop this! \_ I wish somebody told me this a few years ago. I am going to go back to Berkeley to beat him up someday.  screwed by kahan. to go back to Berkeley to beat him up someday. "You are stupid" was what he told a friend of mine who worked for him.  screwed by kahan. \_ Sit through it and hope you gain clue some day. \_ realize that computer science is that to which math 55 is relevant. \_ Hm, good point. 
2001/11/89 [Computer/Theory] UID:22971 Activity:very high 
11/7 I hate math 55. If CS is an extension of math 55, I'm gonna drop out. \_ if you don't like math 55 you are better off doing something else with your life. You will not be happy as a programmer. \_ I don't think math55 is a good measure of what a career in cs is like. \_ career in CS or a career in programming? they are not the same thing. \_ *sigh* \_ well, obviously if you just want to be a code monkey, you don't need to be able to grok math55. hell, you don't need to go to college either. \_ What's your definition of a code monkey? \_ me.want(obBanana);... \_ well, if you have a million monkeys sitting in front of a million terminals for a million years.. \_ You end up with many dead monkeys? \_ UCB undergrad CS labs? \_ Definitely, most CS majors have no pure theoretical interests, they are looking for a well paying job; math 55 will definitely help sharpen certain mental attributes, but most of CS is long work at the lab and understanding of the theory in 170 series.. \_ false dichotomy: pure theoretical interest vs. looking for well paying job. there are many cs research areas that are not theory and some (not many) theorytype jobs that pay well. \_ Well, I think the really smart ones go out and get high paying theory type jobs. The fact of the matter is that the average CS graduate (undergrad) is not going to be the next brilliant theory head, and will just want a good salary and a good work enivironment. Oh yes, btw, I think you have to be absolutely brilliant to do theory type jobs, I dont think most people fall into that catgeory either. \_ The really smart ones format their motd entries correctly. \_ there are certainly CS areas that don't need as much discrete math, e.g. AI, information modelling, or networking even. but if you hate continuous math (calculus, series, etc) too, then give up now! any heavy CS requires a math strength _somewhere_.. \_ Um, graph theory is used pretty heavily in the above disciplines, and Math 55 is a great introduction to graph theory. \_ AI doesn't need discrete math? I see... \_ Well... it depends how intelligent you want it to be. \_ no, it doesn't. AI needs as much combinatorics, graph theory, and numerics as any other aspect of CS. you're deluded. if you don't understand math 55, you don't understand algorithm. if you don't understand algorithm, you're doing something else. and you certainly don't need to know what they teach you in university to do those things. \_ that was sarcasm, but thanks for the rant. \_ I feel for ya. I hated Discrete Math as well (even though I didn't take it at Cal). There're many things in life that won't require Discrete Math and don't let that intimidate you from pursuing CS. Besides (this is going to totally invite spam), you don't always get to work in fields that you majored in. Most people would take a job in fields that they're "trained" on; whether trained in school, internship, etc. AFAIK, system administration doesn't require Discrete Math. But at the same time, when you commit yourself to CS at Berkeley, you're setting yourself up for research, software/system development or teaching. 
2001/10/30 [Computer/Theory] UID:22869 Activity:high 
10/29 ssh cipher: blowfish or 3des... and why? \_ 3des is more processor intensive. Blowfish is younger, but theoretically superior because of some math that I don't understand. \_ Blowfish. Its faster than 3des and can support a longer key length (448 bits vs 168 bits). There are no known exploitable weaknesses in full Blowfish (Blowfish with reduced rounds is exploitable). \_ The key length difference between two _different_ algorithms is irrelevant unless you have a thorough knowledge of the actual keyspace _AND_ decryption complexity of each, which we don't. alexf \_ Blowfish. It's faster and as (if not more secure) as far as anyone knows. If you're securing stuff against a hostile gvt, perhaps you should stick with something safer and slower, otherwise don't kid yourself, blowfish is fine. \_ "Make no mistake! We will crack it for surveillance purposes!" \_ muahahahahh! your puny crypto will fall before our mighty quantum computer! ...just give us another few decades. a physicist \_ wouldn't this lead to an unbreakable cipher as well? or is this only if you have a quantum comp also? \_ Comrade, I am having zee unbreakable cipher. It iz called ze onetimepad. We have been uzink it for years you unwashed american pig dog. 
2001/9/21 [Computer/SW/Languages/C_Cplusplus, Computer/Theory] UID:22569 Activity:nil 
9/20 How do I pass a pointer to a member function in C++? I'm trying to use mergesort (stdlib.h) like this mergesort(array, arrLength, sizeof(int), intCompare); where intCompare is my comparison function but I'd like to put intCompare into a class. But mergesort(array, arrLength, sizeof(int), MyClass::intCompare); doesn't work. How do I do this? \_ You can't. \_ You can't because a member function is not quite a function in a sense you are used to. You see, all member functions implicitly assume a 'zeroth' argument, which is always the object on which this member function is invoked. In general, you do not know the identity of this object at compile time, and even if you did, there is really no syntax in C++ to allow you to specify this information in the function pointer type. What you want is to create a wrapper function, which calls the member function on an appropriate object. Then you can pass the pointer to the wrapper function to your mergesort routine. It's worth noting that object oriented languages with closures and curried functions, like ocaml, avoid this problem altogether.  ilyas 
2001/9/19 [Computer/Theory] UID:22535 Activity:nil 
9/19 What's the difference between a computer geek and a computer nerd? \_ nothing, really. the terms have no bearing on intelligence, so don't bother to whip up any "geek/nerd pride" 
2001/9/1819 [Computer/Theory] UID:22512 Activity:very high 
9/18 I never took 172 and don't have the Sipser book. What is a good alternative? \_ the files in <DEAD>wwwinst/~cs172<DEAD> are protected. Does anyone have the files where I can download them from? Thanks! \_ Papadimitrou's book is very good. But Sipser's book is not bad either  just buy it.  misha. \_ in what respect is it better than Sipser? Papadimitrou has two books, one is Elements of the Theory of Computation and the other one is Computational Complexity. Which one are you refering to? \_ "To which are you referring?" \_ "To which do you refer?" \_ "What book is you tawkin' 'bout Huck?" \_ A good alternative to not having this book is to buy this book... or check it out of the library. \_ Nerds usually like "hopcroft and ullman" it's sufficiently statusgenerating. \_ in what respect is it better than Sipser? 
2001/9/1516 [Computer/Theory] UID:22448 Activity:high 
9/14 I'm trying to buy the 2nd edition of Intro to Algorithms from mitpress, but they're telling me it can't be sold in the US. Does anybody else know about this? They're not being overly helpful. \_ nothing about it on the books pages on http://www.amazon.com link:www.mitpress.com or http://theory.lcs.mit.edu/~clr You sure they just didn't mean it's not printed yet? \_ Cody's books has copies of the 2nd edition  twohey \_ What's the difference between the two editions? \_ Cody's Books has the 2nd edition in stock. twohey 
2001/9/6 [Computer/Theory, Academia/Berkeley/CSUA/Troll/TJB, Politics] UID:22335 Activity:nil 
9/6 Rank in descending order of intelligence: alexf, misha, mconst, tjb, twohey 
2001/8/19 [Computer/Theory, Uncategorized/Spanish] UID:22173 Activity:nil 
8/20 I'm having a hard time figuring out some of the math behind diffehellman. The "key", pardon the pun, to the security of dh seems to be is that calculating g**(x*y) mod n is hard for an attacker. Say x and y are small, less than 1024, couldn't an attacker just precompute all possible K (and K') based on n and g and then do a table lookup based on the X and Y values that are exchanged to get K (or K')? It seems to me that you need to make x and y quite large to have any sort of confidence that K and K' are secure. BUT, a large x and y means the compute time per key exchange is exteremely long. What are commonly used upper bounds on x and y? Sorry, if my questions are exteremely simple, I don't have a good handle on this material yet. 
2001/8/16 [Computer/Theory, Computer/SW/Unix] UID:22139 Activity:low 
8/16 http://www.bustybrainybabes.com/tour/intro.mpg \_ woah! \_ Uhm...that's sort of scary. \_ Yucks! \_ She can't be THAT brainy  she still uses telnet. 
2001/8/2 [Computer/Theory] UID:21985 Activity:very high 
8/2 What does "x (mod y)" in mathematics mean? \_ how are you a computer science undergraduate without knowing this? it's the remainder of x divided by y. \_ Use the distributive property: x (mod y) = (x mod) (x y) Makes sense now? \_ I still don't get it. What does the latter mean? \_ Man, I was kidding. Sorry about that. Seriously, though, "mod" just means the remainder, like from elementary school. So for example, 4 mod 3 is 1. \_ You are missing the minor technical distinction between "x mod y" and "... (mod y)", see below. alexf \_ Well there was no way for me to tell he was referring to FLT until he brought it up. That would have changed my explanation. \_ But that's not the theorem. It shouldn't be an = sign, but rather a "congruent" symbol, which is 3 horizontal lines. It'd take a little too much time to explain, so I'd pick up a Math 55 text for more details. \_ Well there was no way for me to tell he was referring to FLT until he brought it up. That would have changed my explanation. \_ You're not being blamed; your explanation is reasonable given that you weren't specifically asked for distinction between mod and (mod). alexf \_ I just read about the Fermat's Little Theorem, which states "Let p be a prime which does not divide the integer a, then a^(p1) = 1 (mod p)". I don't understand what the "1 (mod p)" part means. Isn't "1 (mod p)" always 1 for all p greater than 1? \_ "(mod p)" after an _comparative statement_ means that that statement is checked only after taking the remainders of both sides when divided by p. So, the following two lines, e.g., are equivalent: a=b (mod p) a mod p = b mod p (Note that to make the different meaning of the "=" clear, the "proper" way to write the first line uses a 3line "=" sign instead of the usual =). alexf 
2001/7/16 [Reference/History/WW2/Germany, Computer/Theory] UID:21806 Activity:nil 
7/16 I'm new to CSUA. Can somebody tell me who the following people are (one sentence per user please thanks): tjb trevor j. buckingham. controversial republican that complains about being oppressed and spouts lots of harsh remarks on newsgroups. smokes a lot of pot and is into turntablism. paolo ??? tom grouchy bike Nazi kchang ??? nweaver (what is "I partied with nweaver?") ilyas russian guy who talks about physics and "AI" and "hard" theoretical problems. ali arrogant Irani who knows C++ well. but no one cares. 
2001/7/11 [Computer/Theory] UID:21762 Activity:nil 
7/10 AI is the stupidest movie I've ever seen. It insults my computer science intelligence. Don't waste yer $ on it. \_ stupider then The Net? that actually sounds kind of like an impressive feat. \_ I actually enjoyed it a lot, up until the the underwater Manhattan scene. They could've and should've ended the movie there. \_ Speaking of movies, go watch Memento, especially if you like puzzles. \_ Good for the first hour, soso for the rest. \_ The maleloverobot is funny. They should've made a female version and a futuristic commercial for http://www.realdoll.com 
2001/6/21 [Computer/Theory, Computer/HW/Drives] UID:21587 Activity:high 
6/20 Robotech DVD's released. Bad ass. Box set available with extra disc featuring the Robotech pilot. \_ fuck that! I'm waiting for the Macross DVD set with the REAL voice actors and non of this "flower of Life" Carl Macek bullshit. \_ Fuck that, I'm waiting for Space Kitteers dvd's and Gaiking. Gaiking Space Dragon on DVD would rock. \_ I am Cyborganizer. \_ cheapURLP \_ #t \_ Robotech? Bad Ass? Surely you jest. \_ Yeah, it only created the giant fighting transforming robot genre. (See, Transformers, Evangelion, to name a few). \_ my ass it did. Maybe you forgot Mazinger Z? Of which Voltes V was the sequel, and then on to Starbirds, and then on to Dynamos... talking 70s here. \_ Create the fighting transforming robot genre? Hardly. Voltron for example was out many years before Robotech. As I recall the same is true with Transformers. If Robotech inspired the creation of the hideously bad Evangelion, well that just another reason that it sucks. \_ man Gundam was out before any of them. Stop watching dubbed anime, and watch the "real" stuff. \_ Quibble over semantics if you must. Macross (later rebroadcast and munged together with two other series as Robotech in the U.S.) predates Voltron, Transformers, and Gundam. Get over it. \_ uh, Macross originally was intended to PARODY the giant robot genre. there was plenty of giantrobot anime well before Macross (Danguard Ace, Grandizer, Gaiking, etc.). \_ Robotech, Harmony Gold, and it's progeny must be destroyed. Makurosu, ai oboeteimasuka, forever. 
2001/6/19 [Computer/SW/Security, Computer/Theory] UID:21573 Activity:high 
6/18 I have a question about diffiehellman. After going through the initial key exchange and generating the session key k', how do you use this key with 3des or blowfish? Do you just trucate the key to the appropriate length (doesn't seem right) or is there some other method? tia. \_ Probably feed the key into a one way hash function (i.e. MD5) that outputs the appropriate number of bits. \_ This is correct. You would use a hash function. However, you should not use DiffieHellman straight, much the same as you should not use plain RSA. Get a cryptography book and read about it. \_ Okay, I understand the bit about the hash function, but I don't understand why the session key k' can't be used directly? I've been referring to Applied Cryptography, but I can't seem to find a place where he explains why the session keys should not be used directly. \_ Here's a handwavy argument: Your DH key must be larger than your 3DES key since otherwise it's easy to break DH. This means that you'll have to shrink your DH key to make your 3DES key. You want to make your 3DES key by using all of the randomness that you've got in your DH key, but you don't know if truncating the DH key will do this. However, you DO know that using a good hash function to make your 3DES key will conserve all of the randomness of your DH key. \_ I guess I wasn't clear. I understand that I need to hash the session key in a way that preserves the randomness of the key and that I need to use the hash value as the key for my crypto algorithm. The bit I don't understand is related to the following: I keep reading that one should use the hashed value of the session key *only* for encrypting a different secret key and then that encrypted secret key should be transmitted so that all other transmissions are encrypted with the secret key rather than the hash of the session key. Why can't I just keep using the hash of the session key? It seems much simpler to do this than to maintain a separate secret key. 
2001/6/1719 [Computer/Theory] UID:21552 Activity:high 86%like:21549 
6/17 [REPOST] Anyone know of a good introduction to the GellMann 8 Fold Way that can be understood by a engineering grad (7 series plus one ud phyics class and one ud math class)? \_ forget it. \_ forget it. There's so much physics out there that is accesible to someone with the background you describe, that will actually teach you something useful about how the world works, and can be applied to useful things like how devices work. why fuck around with useless stuff you need grad level math and physics to understand? \_ Because it is interesting. \_ No. AI is interesting. \_ I've worked on AIs before (GA,NN,ES,FL) and IMHO physics is much more interesting because it is understandable via mathematics. AI is just a bunch of weeny structured programmers whinning about lisp, scheme, semantics, cognition, determinism, etc without ever producing anything useful because the concepts of intelligence and selfawareness are too hard to reduce into programmable abstractions. About the only promising avenue of research in AI is machine learning, but once you get right down to it, the most fancy machine we have today is about as smart as a fungus or possibly a protozoan. \_ You are a stupid, stupid troll, friend. You need to read a real AI textbook. What you are whining about is not AI, it's cognitive science. \_ Pick up any intro abstract algebra or group theory text, read up on basic group theory, read the stuff on SP/SU groups, and then we talk. From http://www.pickupguide.com *Warning*, DO NOT try and develop a relationship with a girl *before* you sleep with her. Because, then your head gets all messed up, when they decide they like that relationship they already got, and don't want to risk ruining it by having sex. \_ Why would you develop a relationship with her after you've slept with her? Maybe it should be: "DO NOT try to develop a relationship with a girl. Sleep with her as often as she'll allow until she stops putting out." Seems like better advice that way. \_ The age that I am, he's the reason I'm still single! These are the guys who are out in the world. \_ You are still single because you are a loser who blames sexually satisfied people for your problems. Women can smell a loser a mile away. \_ Is that why they avoid me? \_ Yes. 
2001/6/1415 [Science/Physics, Computer/Theory] UID:21510 Activity:kinda low 
6/13 Long lived quantum entanglement of 2 macroscopic objects has been achieved: http://xxx.lanl.gov/abs/quantph/0106057 Food for thought.  ilyas \_ Thaumaturgy scheduled for teaching at UCB in year 2011 \_ We will teleport your gonads into the icy environs of deep space! \_ Eh. Note that this has only been submitted to Nature and thus hasn't passed peer review yet. The results they claim do sound impressive, but I'm holding off on the party for time being (and I'm not nearly enough of a physicist to evaluate actual procedure used and the consequent claims' validity). alexf \_ .5 milli seconds is longlived? not useful for engineers yet.. \_ for quantum computation, that is wuite long. If you could get that kind of lifetime for a 10,000 qubit system, you would have a real quantum computer. \_ Even 200300 qubits at that speed should kick the crap out of the current state of classical machines. alexf \_ 0.5 ms is long for all scientist/engineers. \_ that's not what their sexual partners think. 
2001/4/2425 [Computer/Theory] UID:21088 Activity:kinda low 
4/24 "NewNew Math": http://dailynews.yahoo.com/h/ipts/20010423/cm/_newnew_math__1.html American kids are so spoiled. \_ In what way does it spoil a child to *not* teach them math when they need it to do well in life? I think this is abuse. \_ they have these adds on the radio where i live trying to convince people that taking alot of math leads to great riches. I keep thinking of the people i know who *really* know alot of math, living on jolt and pop tarts and sleeping on matresses in Evans hall. \_ There's a difference between "most kids need math to do well" and "needs lithium to do well". 
2001/4/18 [Computer/SW/RevisionControl, Computer/Theory] UID:21018 Activity:moderate 
4/17 Lookin' for something that can generate a dependency graph for C's #include and Makefile's include. Thanks! \_ Check out AST on freshmeat or sourceforge. It's out of AT&T research. \_ ast.pl is also good. \_ make depend? 
2001/4/16 [Computer/Theory] UID:20992 Activity:high 
4/15 Anyone know anything about the profs teaching Math 110 in the Fall? Specifically, Anshelevich, Givental, Harrison, Neu, and Pugh. Plus, "The Staff". \_ You should really wait till either Demmel or Kahan teaches 110 to take it. \_ Harrison was the worst prof. I had at Berkeley. Avoid like the plague. \_ Pugh is very good, though he gives difficult exams. \_ Is he related to "plugh", "xyzzy" or "say echo"? \_ Both Givental and Pugh have a rather sizeable "hardcore" reputation attached to them. Which, of course, may be good or bad, depending on what you're looking for. \_ "The Staff" is long and wooden. It's currently hidden in Wu's office. On occasion, it is inserted up naughty students' recti for maximal educational effect. \ you know for the longest time i misparsed this thinking it was a reference to prof. wooden and some new prof long ... until i remembered it was woodIN@math ...heh ... speaking of H. Woodin, the tarski lectures begin tomorrow on set theory foundations. http://www.math.berkeley.edu/Tarski/2001tarski.html \ I would take Pugh, i dont know Anshelevich. Pugh used to give \ this guy seems to have a good sense of humor. "open problems" in 214, so he has his sick bastard side but he does care about teaching. I would avoid his wife [Harrison]. ok tnx. \_ hi psb! (note: psb supports this guy. once again, testament to him being hardcore. it's up to you to decide whether this is actually a good thing. which it may or may not be.) \_ Only GPA matters. \_ I can recommend Neu. Homework was difficult, lectures were informative, and were tests not very hard at all. informative, and tests were not very hard at all. \_ I had Neu for 128A. I concur, although his tests weren't exactly cake. They just seemed like it after the killer homework. dim 
2001/3/1920 [Computer/Theory] UID:20838 Activity:high 
3/18 Exteremely intruguing paper (as seen on wall.log): http://www.cs.umaine.edu/~chaitin/cmu.html \_ snooze... \_ take your philistine ass back to where you came from \_ Yeah, get your ass back to High School, you aren't mature enough to be in college yet. \_ That algorithmic complexity/information theory stuff is more religion than math or science. Godel and Turing's results are far more practical, useful, and valuable. \_ Advanced mathematics is just welfare the insane. \_ "Chaitin's Omega: * irrational number * incompressible number  Oliver King * all around fucked up number"  Crow T. Robot \_ Oh shit! It is now proven that we can't possibly prove everything. (Seriously) \_ So is it soon going to be deemed unconstitutional to ask our troubled high school students to prove something in final exams? (Just like how teachers are not allowed to teach Darwin's Theory or Evolution, or something like that.) \_ Huh? High school finals? It's all about feeling good. They don't have to prove anything but that they've been properly socialised and turned into good little robots. \_ Intruguing? Paper? It was interesting, yes, but hardly Intruguing. It wasn't a paper. It was a lecture. \_ Okay, I got the paper part wrong. \_ I think that people who are underwhelmed by this do not understand the consequences of "maximally unknowable" within the context of a proof system. 
2001/3/1415 [Computer/Theory] UID:20788 Activity:nil 
3/14 Did anybody record the Anne Murray PBS special last night on PBS KCSM? 
2001/2/911 [Computer/Theory] UID:20547 Activity:high 
2/8 I have 20 positive numbers and five desired totals (all floating point numbers). I need to pick five mutually exclusive sets (not necessarily exhausive) of numbers from the 20 numbers, and then sum up the numbers in the five sets to get five totals. My goal is to minimize the meanofsquares of the differences between the five totals and the five desired totals. I wrote a C program to do the bruteforce tryall combinations method (that's (5+1)^20 combinations), and it looks like it's going to need ~120 years of CPU time on my PII 350MHz! Is there any better algorithm for doing this kind of things? Thanks.  yuen \_ why isn't it 20!/(205)! ? \_ This is known as a 'bin packing problem.' It is NPhard (in other words it is at least as hard as any problem in NP).  ilyas \_ exclusive means (20! / (5! * (205)!)) right? \_ No, that would be the # of combinations to pick one set of five numbers from the 20 numbers. But in my case I need to pick five exclusive (ie. nonoverlapping) sets instead of one, and each set can contain anywhere from zero to 20 numbers.  yuen \_ Can you use an approximate solution? I think of this problem as being more similar to integer programming. It's equivalent to min_A Axb^2, such that 1^T A <= 1, where x is your vector of 20 positive numbers, b is the five desired sum, and A is a 5x20 matrix with the entries constrained to be either zero or 1. You won't be able to reach the exact global optimum in poly time, but perhaps you can try adding an L_1 constraint to the objective? i.e. min_A Axb^2 + lambda*sum_ij A_ij, s.t. 1^T A <= 1 L_1 penalty punishes any nonzero values, so A will have as many nonzero values as possible. Of course you're still left with the problem of deciding which entries of A to clamp to 1... Uh, need more optimization fu.  alice \_ If my memory serves me right, LP relaxation isn't all that great for binpacking. The first reference point to check would be _Approximation_Algorithms_For_NPHard_Problems_, edited by Dorit Hochbaum: http://www.ieor.berkeley.edu/~hochbaum/html/bookaanp.html It should be available at any university library by now. Chapter 2, written by Coffman, Garey, and Johnson (yes, THOSE Garey and Johnson), is "Approximation Algorithms for Bin Packing: A Survey" alexf \_ I'm not up to date with NP approximation algorithms, but LP might be the easiest (and most practical) to implement than more complicated methods. BTW, one can express 0/1 constraints using (2A_ij1)^2 = 1. The problem then is no longer LP, but you can get a lower bound using Lagrange relaxation (i.e. switching min & max), then use gradient descent or something on the lower bound function, which is usually much more tame than the original problem. This all sounds very complicated but is actually doable. I just don't know how good the approximation is. Maybe the book will say something about it.  alice \_ Thanks for all the suggestions. Understanding what you all said is already hard enough. Gotta refresh my 170series knowledge first before I go from here.  yuen \_ can you think of a useful lower bound for what you're suggesting? you are talking about lower bounding Axb^2 + l*(2A1^21) right? I'd bet that your regularization term smooths out the surface so much that you could just solve this using gradient descent. that you could just solve this using newton raphson. \_ this is one smart broad. 
2001/1/8 [Computer/Theory] UID:20269 Activity:nil 
1/9 How do I pass by reference in matlab/octave? Also is there a better alternative for numerical computing? I've heard that the numerical package for python is decent. Any comments? Thanks. emin 
2001/1/34 [Computer/SW/Mail, Computer/SW/Security, Computer/Theory] UID:20228 Activity:nil 
1/2 I've been getting the following error message repeatedly lately. The authenticity of host 'quasar.cs.berkeley.edu' can't be established. RSA key fingerprint is 14:1f:b3:63:83:6a:fe:73:4e:fa:64:30:9c:9f:c3:c8. Is this a problem w/ quasar or is it the soda ssh client? Why doesn't it allow me to add quasar to my list of trusted hosts? 
2000/12/2327 [Science/Disaster, Politics/Domestic/President/Bush, Computer/Theory] UID:20165 Activity:nil 
12/22 Is the math instructor "Hurricane" Wu mentioned below the same as the famed but controversial human right activist Harry (Hongda) Wu? I heard the latter once was at Berkeley. \_ yes \_ no \_ No. Prof. Wu from the math dept. isn't human and therefore has no interest in human rights. \_ Harry aka Hurricane used to work at Los Alamos until he was fired for mishandling his tapes. \_ Same Wu? The Joy of Lecturing by H. Wu http://www.math.berkeley.edu/~wu/lecture.pdf \_ The last man on the planet to write about better teaching. 
2000/12/2225 [Computer/SW/Languages/Java, Computer/Theory] UID:20156 Activity:moderate 
12/21 Turns out that I got a lower grade than I expected. If enough people in the class ask for a recount (all the grades), will it help? Has it ever happened before? \_ This doesn't fall under the jurisdiction of the Florida Supreme Court, but then again, that never stopped them before. Go for it! Every grade must count! Count all the grades! Even the smudge you made on the paper in the place where the correct answer might have been implies you had "studentintent" to fill in the correct answer. \_ i had a teacher in hs who used to drink wine while he was grading and get really wasted. sometimes he would only grade half the exam and you would get a 20 out of 100 or something, and demand a recount and get a 90. suprisingly, he was a really good teacher otherwise. \_ Just don't let the professor stall till the grades are certified as final. \_ Grades may only be changed to correct errors, not because the \_ You can also file an appeal to a grade with some dept. at Cal. It doesn't only have to do with errors. professors grading scale was too high. (However, if the prof consistently gives too many bad grades, he will get in trouble eventually  see the case of Prof. Wu & Math 1A/1B in the early 90's.) \_ yah, that foolio got fucked up. He's now been demoted really awful in math. \_ are those errors as in "Ooops, I read that 3 as a 5 and misgraded you" or as in "Oh, I guess you answer is correct mine (prof's) is wrong" \_ I remember a horde of people dropping math 1a from wu in 91 \_ I had Wu and I think I got a B. What was his deal? \ this is pretty funny ... psb http://www.jiggscasey.com/slappy/book_of_wu.html \_ excellent linkage, thanks psb. READ THIS people. \_ it isn't *that* funny \_ Yeah, did something happen to Prof. "Hurricane" Wu? \_ As I recall, he was still allowed to teach but was no longer allowed to have any part in the grading process including setting the standards, creating the curve, grade appeals, etc. That was all given to his head TA. Me? I was there for 10 minutes of lecture 1 and walked out. I knew what I was looking at. This was while he was still allowed to destroy hapless freshies. Any survivors from Prof Shang's math 1b class still around? Oscillating circles. 
2000/12/17 [Computer/Theory, Computer/SW/Compilers] UID:20114 Activity:high 
12/14 I totally don't see any relevance in teaching Lambda calculus in an undergrad course to describe stupid semantic rules. Fuck Aiken and his stupid ambiguous exams!!! researcher) \_ usually to emphasize its distinction from an algebra. keep in mind that "the calculus" taught in high school is really "the calculus of infinitesimals." \_ this is a simple question but: why do they call Lambda calculus "calculus"? \_ It's a way of calculating stuff. and it sounds cool (to a researcher) \_ usually to emphasize its distinction from an algebra. keep in mind that "the calculus" taught in high school is really "the calculus of infinitesimals." \_ I actually wish Hilfinger covered some of that stuff. \_ Hellfinger just asked you completely unrelated questions pertaining to English lit or what happened on a certain date in France that never existed. Often times those answers were as simple as zero or nothing. Nevertheless, I enjoyed pushing the very limits of my parasympathetic nervous system. \_ About dates that never existed? Then the answer is pretty obvious: nothing. \_ Has a Hellfinger exam question list been compiled? I know two: one about some british poem and another about some Spanish/Mexican revolutionary figure. And i think it was french lit and english date. \_ "someone" should definately do this... I remember two: what was the HMS Java? I don't remember the answer but can look it up. The other was a line from "Ode on a Grecian Ern". If more people post them, I can compile the list. \_ this is a good idea. A question from one of my exams was "Where do the poppies blow, between the crosses, row on row?" (A: In Flanders Field) \_ If you're going to teach language semantics, at all, there are two ways to do it. (a) Operational semantics, in which your language description is written in something that works like Tcl. This is icky and went out of vogue in the 70s. (b) Denotational semantics, in which your language description is written in something that works like Scheme (the Lambda calculus). Thank God that your undergrad compiler class is not about stupid shit like how to write a lexical analyzer, like mine was. Also, http://mw.com's definition of "calculus": 1 a : a method of computation or calculation in a special notation (as of logic or symbolic logic) If semantic rules are stupid, then what exactly is a programming language? blojo \_ Programming languages are hard. Let's go shopping! \_ lexical analyzers are not "stupid shit". especially the practical applications of 172 stuff. and neither are lambda calculus or type systems. how can you justify asserting that they are? this whole thread is just bizarre. \_ well, they're kind of overkill in general when all you really need is an sexpr reader. \_ How quaint; how '50s. Dude, humans were not made to read sexpressions. \_ I wasn't made. I was born. human \_ A lexical analyzer is a necessary component of a compiler and it's worth maybe spending a week talking about them. It is NOT worth spending 1/3 to 1/2 a semester talking about them and doing major projects regarding them. Any programmer with a clue can write a lexical analyzer. It just takes general programming knowledge. The important knowledge in compilers, the domainspecific stuff that the class should be spent teaching, is all about semantic analysis, code motion, and maybe provability. blojo \_ Um. I can write a lexer very quickly with automated tools like Lex. Very quickly == hours, not days. \_ Yeah, that's like my point, see. blojo \_ lexical analysers are good for building language > machine translators, but the underlying theory is useful as a basic model of computation. knowing all the DFA, NFA, REGEXP equivalence shit is really useful if you do CS. also, there is a lot of theory that goes behind building YACC. you could either use YACC as a customer (as in the way compiler writers do) or you could delve into the theory, like the way language people do. i think you're a lazy bitch if you think this shit is bunk. ali \_ Bunk! Bunk! 
2000/12/67 [Computer/SW/Languages/Misc, Computer/Theory] UID:20012 Activity:high 
4/249 I'm interested in treepattern matching algorithms. I'd like to implement a language independent, synthesized and inherited abstract syntax tree comparitor. Where is a good place to start? \_ well, I actually am looking into the same stuff, this work can do fuzzy tree matching, giving you replacements, renaming, insertions, etc. http://www.cis.njit.edu/~discdb/treematcher.html \_ probably not here. \_ http://www.google.com/search?q=tree+pattern+matching+algorithms \_ Check out the XML tree diff from IBM. Not sure if there is code in PD, but they at least quote source references. < O(n^4) 
2000/10/3011/1 [Computer/Theory] UID:19603 Activity:nil 
10/31 http://www.geocities.com/st_busygin/clipat.html \_ so is this under peer review 
2000/10/20 [Computer/Theory] UID:19531 Activity:very high 
10/20 Except for Graphics, it seems like high level math doesn't seem to be needed in computer science. \_ heh heh. heh heh heh. heh. \_ yes. \_ except for graphics and for anything that involves fucking modeling. ali \_ well there's that whole crypto thing... \_ I want to fuck a model. Does that require high level math? \_ no.  paolo. \_ in fact, having highlevel math would probably hurt your chances. \_ yes. math alum \_ this is bullshit. it's true that math major can't score, but well positioned math skills always help you score. chicks dig guys who do math. it completes them. it's a well known fact. ali. \_ if you include abstract algebra, then it is a big help in compiler implementation, language design, AI language processing... if you include topology and/or graph theory, then it is useful all over the place. there are more examples but the motd isn't big enough. \_ Abstract algebra for 'AI language processing'? Where are you pulling this shit out of? Idiot. Go learn AI. \_ The motd is infinite!!! 
2000/10/1819 [Computer/Theory] UID:19514 Activity:very high 
10/18 Anyone here take any actuarial exams? Anyone know what its like to be an actuary? I'm graduating with a applied math degree (CS) next year and I hear actuary's make decent money. \_ what does the CS in CSUA stand for? \_ Crazy Shitheads \_ Criminally Supercilious? \_ Cum Suckers \_ Cock suckers \_ Cunt su... no, it can't be that. \_ Computer Science \_ It can't be that. \_ Commuter Science \_ Cognative Science \_ or if you ever could spell: "Cognitive Science" \_ Completely Silly \_ Casual Sex > Cow Sex \_ Cow Shit \_ Creaky Sysadmins \_ Cal & Stanford \_ Clitoris Studs. 
2000/8/1516 [Computer/Theory] UID:18997 Activity:very high 
8/15 Breakthrough in quantum computers! http://biz.yahoo.com/rf/000815/n15481726.html \_ Fixed the link. \_ So it can solve a particular problem in one step instead of four. Isn't that the result of a change in algorithm instead of a change in the underlying electronics? How does switching to using atoms to implement the logic make it possible? \_ No, it is a change in the underlying electronics. Quantum algorithms can't be done on nonquantum computers.  mikeym \_ The fact that they successfully demoed this essentially confirms that it is indeed physically plausible to implement the Quantum Fourier Transform, which is at the heart of, among other things, Shor's polytime quantum integer factoring algorithm. Presumably what polytime factoring will do to computing needn't be explained. Here's some reading for details: http://www.cs.berkeley.edu/~vazirani/qc.html http://xxx.lanl.gov/ps/quantph/9809016 alexf \_ dude, I work in the industry, I don't understand this crap. Should I go back to school? \_ if you wanna work for peanuts as an academia phreak, YES \_ if you want to understand it and/or contribute to it, YES. if you don't, don't bother. it's rather unlikely that there'll be practical industry demand for quantum coders before you retire (or at least are too old to pick up an entirely new paradigm). \_ just wait for phillip's QC API oj \_ Too old? Like, his brain would seize up and he'd be incapable of learning something new? All his bits are in use? Unlikely. \_ When you're 60, you will, in all likelihood, not be to willing to learn a whole new world from scratch \_ http://www.sciam.com/1998/0698issue/0698gershenfeld.html This is a Sci Am article on the topic. For people who haven't had Quantum Electro Dynamics in college (which is like 99.999% of the world. :) \_ See second link above (LANL); it's perfectly accessible for anyone smart enough to log in to soda and read the motd (ie not that smart). \_ hey1! :( 
2000/8/1415 [Computer/Theory] UID:18977 Activity:high 
8/13 Is there such a thing as an algorithm that can output another algorithm? Kind of like self tuning, self evolving algorithm? \_ Yes. Not as glamorous as it sounds; see also: genetic algorithms, Remez algorithms, datadirected programming \_ bison, yacc, (f)lex, and many many more. not self tuning though. \_ There are algorithmns that can learn from data. It's really not as sexy as some people seem to think, just an application of statistics. \_ selfpropagating neural networks \_ Neural nets are data structures not algorithms. And there is no such thing as a selfpropagating neural net. There are neural nets with feedback loops, but they don't output 'another algorithm.' They have some uses in time \_ Tell us of the stars.... series prediction. Please don't be an idiot, read Russell's book or something. \_"A neural network is a processing device, either an algorithm or actual hardware, whose design was motivated by the design and functioning of human brains and components thereof" from the NN FAQ. And I believe it is theoretically possible for a NN to output another algorithm. Sign your posts. williamc \_ A NN FAQ is not an AI authority. For one, this "FAQ" seems to imply that an algorithm is a processing device, which is idiotic. And a neural network cannot 'theoretically' output an algorithm. It outputs an array or list of floating point numbers. I recommend Christopher Bishop's neural networks book as an excellent way of learning more. \_ It's easier to babble on the motd and toss buzzwords around. \_ how dare you call me a "data structure". brain \_ feedback loops \_ Does selfmodifying code count? \_ Optimising compilers fit, too... \_ and my thesis! 
2000/8/34 [Reference/Military, Computer/Theory] UID:18857 Activity:nil 
8/2 http://www.faisal.com/geek/guns.html features jordan hubbard, while still at berkeley. 
2000/7/29 [Computer/Theory] UID:18814 Activity:high 
7/28 Rest In Peace, John Tukey \_ I don't know who this guy is, but doesn't he deserve the bits required to spell out RIP? \John Tukey was one of the most important mathematical statisticians "ever" ... spent a long career at Princeton. Long list of results and many many important collaborations in many fields. First rank giant in 20th cent mathematics. psb \_ He was one of the people who rediscovered the Fast Fourier Transform algorithm and popularized it's use in Digital Signal Processing. \_ that's nice, but did he join a startup? Did he get XXXX options? Is he a millionaire? I am not a rocket scientist but I know how to invest and work at the right place. I'm a twentysomething multimillionaire already. Success is measured by your $, not that IQ or humanity bullshit. Fuck academia. Industry rewls. industry pimp \_ Twit imposing their personal morals on others removed. My personal morals involve censoring others when I feel what they said was distateful to me just because I can, just the same as you've done. Let's hear it for more censorship in academia. 
2000/7/17 [Computer/Theory] UID:18700 Activity:very high 80%like:18708 
7/17 What is the operating principle behind Computer Associate's "Neugents" and why is it so damn smart? \_ they took a neural imprint of ted nugent \_ It's all marketspeak. They are using straightforward neural networks, a technology that has been available since the late 60s.  ilyas \_ No no no no no, the commercial said their new technology was so smart computers can think!!!!!!! \_ Yes yes yes!!! It's just like that computer 'HAL' on 2001! "Dave, what are you doing?/" Ohmigod!!! 
2000/7/1718 [Science, Computer/Theory] UID:18699 Activity:very high 
7/17 i think micro$oft's new language should be called C to match the grade i got in cs164  SocialScience Major \_ WTF were you taking cs164 for as a fuzzy major? \_ To take space away from an EECS major? BTW I didnt really get a C, but it's a better post that way. The major is only fuzzy because Math, CS and Bio havent figured it out yet. \_ Anything that needs to add the word "Science" to it's description probably isn't. Yes, including computer science. cognitive science alum \_ Amen to that. Computer science is either math or engineering, depending on which side you're on. And cognitive science, is, of course, bullshit. \_ No, Cog Sci is either linguistics or neuro anatomy. Don't judge it from the cogsci 5 course you took. cog sci alum \_ I am judging from CogSci100, CogSci101, and CogSci110 courses I took. Linguistics is all seem to warship at berkeley? either bullshit or linguistics. 110/115 series or the 130's. Neuro anatomy is neuro anatomy and those who call it cogsci are simply trying to cover the rest of cogsci's bare ass.  original poster \_ oh, yeah? which of those is the idioitic arguments of that antiAI crusader who you all seem to worship at berkeley? \_ Wow, and this wasn't even me.  ilyas \_ I don't "warship" John Searle. He's a pompous arrogant ass who has nothing to do with anything. cog sci alum \_ I don't like John Searle NOR George Lakoff, and yet I loved being CogSci at CAL.  Marco \_ At least Lakoff wasn't an asshole like Searle even if you don't like his lecture materials. What'd you not like about him? \_ warship? We have a warship? Let's put our football team on it and sail to Palo Alto. \_ We have a speeling disfunkshun. 
2000/7/23 [Computer/Theory] UID:18576 Activity:high 
7/1 Protein folding problem as special case of knapsack. If you're serously interested in the below thread look into it. Otherwise shut up bioinformatics wannabes. bioinformatics stud \_ dynamic programming approximations do DAMN well, though  computer science stud \_ approximations in one domain don't carry over into other domains under most incarnations of mapping. !stud \_ think smithwaterman algorithm (DP algorithm for PSM :) \_ urlP \_ urlP / referenceP \_ googleP \_ what the FUCK are you guys talking about? English motherfucker do you speak it?! \_ Better than you. 
2000/6/307/3 [Science/Biology, Computer/Theory] UID:18573 Activity:very high 
6/30 Now that the human genome appears to be all but decoded. Is there any method to measure the number of bits that are encoded in the genome. IE how does it compare to a modern operating system. \_ Well, the encoding system is using a power of two, so there is a very easy conversion. The problem is that it's not always easy to see where code ends and garbage begins in DNA. \_ ONE HUMAN ~ 4 TERABYTES \_ Uh, I don't have my biochem text with me (on vacation), but I seem to recall the human genome being 2,000,000 kbp (kilobase pairs), or 4 Gbits of data (2 bits/bp). nweaver \_ That's just the program text. The Interesting Question(tm) is How much does it take at runtime? \_ The number is actually much less than that, since the bitstring is EXTREMELY structured. Which means less bits. If I were to guess, you're off by a factor of 1001000. Maybe worse. That doesn't mean anything however, since we know next to nothing about the structure, and won't for quite a while \_ ONE HUMAN ~ 4 TERABYTES NO COMPRESSION, PUNY HUMAN \_ once the genome is there, the interesting stuff begins. For the next 3050 years, I think scientists will be working on the grand \_ try 300500; popular press is just listening to what the funding proposals are babbling; anyone actually writing them makes sure the timespan predicted is long enough "so that i won't be around to be held responsible" but short enough as to not to discourage investment. sad, but true. unification theory of DNA. A physical/biology/mathetical model of the interaction of the different genes. Imagine running a simulation of a new lifeform created by artifically pieceing different genes! The complexity of such a simulation is beyond anything we've done. Today's supercomputers used to simulate nuclear explosions will look like toys next to computers simulating artificial lifeforms. Who wants to guess on the computational power needed to run a simulation of a single cell? \_This is the typical clueless CompSci answer to biochemical problems. I remember once one of my advisors said that the problem with working with computer scientists on biological simulation in actual living cells. Just pick your favority problems was that they just didn't get it. I guess he had a point. Why waste your time trying simulate a complete cell at such a granular level on a computer? We can simply run the simulation in actual living cells. \_ Why bother running simulations of rockets, and atomic bombs? Oh yeah, that's right, if you find something *really* interesting, **THOUSANDS/MILLIONS** OF PEOPLE **DIE**. Apparrently, its true that those who can't do, teach. \_ What are you trying to say? This makes no sense. Just pick your favorite organism and transform them. DNA is cheap and plentiful to reproduce with a little lambda phage, plasmid, and PCR. Also, simulation of a single cell, albeit interesting, isn't exactly \_ >80 column idiocy fixed. Get a clue. tom completely useful. Since we are mainly interested in multicellular organisms, a simulation of intercellular interactions would be much more valuable. i.e. what exactly is involved in the complex interaction of cell signalling during embryonic growth, and how that interrelates to differentiated cells. A more realistic goal is to use pattern recognition techniques to be able to predict tertiary/quarternary structure of proteins and enzymes from DNA, and probably one which is much more profitable than trying to simulate organisms when the actual organisms can be produced cheaply. Go buy yourself a copy of Maniatis. williamc. \_ If you take a pure scientific view, there is lots of value to understanding how cellular processes work, and being able to model them means a huge step toward fully understanding the schemes (algorithms if you will) nature has come up with. From a practical viewpoint, you want to be able to model a cell so you can design your own cellular signalling pathways What you're saying, William, is that there is no value in understanding the inner working of cells, that nuclear transport, mRNA regulation, vessicle trafficking is not impt. Thats a very narrow minded view. \_ What he's saying is that full simulation is infeasible, and suggesting a viable alternative. Get a clue. \_ see below \_ More than Moore's law can produce for you even if it lasts through 2500 A.D.. Without a new computational paradigm, or a better abstraction than sheer chemistry, this will not be practical (in all likelihood) until well past the predicted lifespan of the Homo sapiens species, or even genus Homo. \_ Dude. Do you realize how LARGE the number current_computational_speeds * 2 ^ (500 / 1.5) is? \_ Yes I do. Do you realize that modeling a physical system on quantum level is considered nonpolytime on a classical computer? And do you realize how many atoms a cell contains? \_ In something like 8 iterations of Moore's Law (12 years) you'll be able to read 4 terabytes (the DNA sequence) into RAM. The rest of the cell structure is simple relative to DNA and doesn't need to be fully modeled. By the time you can read DNA into RAM, processors will be running at 256 Ghz, with who knows how many instructions per cycle. That's far more processing power than a cell has. The only computational barrier at that point will be writing the code to model it correctly; that's hard for a cell and much harder for a full organism. tom \_ "The rest of the cell structure is simple relative to DNA"? Get a clue, cs boy. You can read the damn bytes into RAM, but you won't know what the fuck to do with them. Predicting "everything" from DNA, or even a small subset of it such as the general protein problem (folding, interaction, binding sites, etc), may easily, to the best of mankind's current knowledge, turn out to be, oh, say, EXPSPACEhard. All your Moore's law ramblings aren't worth crap until we know SOME fully encapsulated localization structure in the problem (be it DNA, protein, life, etc). Which doesn't seem too plausible. \_ It would be stupid and _unnecessary_ to model the individual atoms to model a cell or dna. For example, weather modeling gets better everyday and they're certainly not modeling every atom in a storm. \_ See above. \_ And do you honestly think we'll still be computing on silicon then? \_ The above was predicated on "no change of paradigm" \_ But can distributed computer help, like what SETI@home does?  yuen \_ Probably not; seti@home relies on the fact that an arbitrarily large amount of computation can be done by any node without needing input from any other ongoing calculations; a cellular model would be much more interactive. Still, I think the assertion that we'll never have enough computing power to model a cell is silly and unfounded. tom \_ 3 words for you  "think avogadro's number" \_ "No one would ever need more than 640k". \_ I have to agree with william. You don't start a computationally intensive calculation at the lowest possible level of understanding. For instance, if you ever want to see a result, you would not start a model of even a modest polypeptide by doing ab intitio calculations on the interactions between individual electrons and nuclei. Modeling an entire cell based on molecular interactions is similarly too complex and really unnecessary. \_I dont understand this fixation with atoms. You dont need to model atoms, just the kinetics and thermodynamics of interactions. Duh, anyone thats knows anything knows theyre not going to figure out interactions in the cell from scratch. We have 100+ years of abstraction to work with. \_ Nobody is talking about simulating cells at the atomic level, dumbass. As for "why not try it on a real cell?" It's a stupid question. It's always more economical to simulate something first rather than try it first. You can change your simulation parameters faster than you can change your realworld experiment. \_ this is utterly false. tom \_ this is the first intelligent thing you've said in this thread, tom How do you think we build cars and airplains and computers? We break it down into components, build models in computers, simulate them, and then build small scale models. Drugs can be synthesized in a computer faster than in real life. I'd love to see how a particular drug will affect a cell even before the drug exit in real life. Science fiction? maybe. But then again, who would have thought of the internet 100 years ago? \_ what is human gnome, and is it better than kde? \_ alot of you missed a point made above, DNA isn't enough! The cell itself carries much info that isn't in the DNA (via already synthed proteins, sugars, biochemical microenvironents, mitochondria and their DNA, imprinting (which the genome project is ignoring), as well as other molecules that we probably don't realize are necessary in a model). Yes, much will be able to be done, but the necessary in a model). \_ the total amount of cell information not contained in DNA is almost certainly less than the amount of information contained in the DNA. So call it 8 terabytes and 13 iterations of Moore's Law. tom Yes, much will be able to be done, but the system will have holes and leave a lot to interpretation. That's not to say that phages, bacterial sims, YACS, . . . are the answer, they also have many, many flaws, but we are getting closer. And it is probably the marriage of the techniques that will produce the answers we are stiving for, with the great aid of human intuition and analytical skills. Anyhow, the 4TB, GB, whatever, of DNA isn't enough. Just imprinting alone would add 2 bits to every base pair (methylated or glycosylated), now add on everything else you forgot to consider. Oh, and don't forget you need the environments of all surrounding systems, i.e. in birth you need the mother, her DNA, and so forth to get it all right. Bottom line, an approximation is better than nothing, but don't get your hopes up too high! \_ The first challenge is simulating an amoeba. tom 
2000/6/19 [Computer/SW/Languages/Java, Computer/Theory] UID:18493 Activity:nil 
6/18 What's the course number of the linear algebra class you need for cs 180 (the computer graphics class)? How hard is it to understand...old eecs alumni tell me it was quite nasty?  no clue \_ Probably Math 54. It's fairly easy if you pay any attention at all (unlike me). Take.. I think.. Borchards if you can. He gives _really_ easy tests. dbushong \_ i thought math 50a was pretty hard. i don't think they explained the concepts well at all. it wasn't until 50b that i started understanding what was going on. 
2000/6/4 [Computer/Theory] UID:18403 Activity:moderate 
6/3 Thank goodness, somebody finally nuked that whole convolution proof thread. Please keep that shit at a math newsgroup and NOT on the motd. \_ I would rather hear about a good proof than about most of the other shit regularly discussed here.  ilyas \_ Tell us of the stars, ilyas.... please lick my hairy scrotum during a prolonged stint during which \_ dear mr. "i would like to hear more about dating vapid playboy bunnies", please lick my hairy scrotum afetr a prolonged stint during which i have not had a chance to shower. Sincerely, Ali. \_ You have a scrotum? \_ you would have licked it sooner had you known? 
2000/6/2 [Computer/Theory] UID:18388 Activity:high 
6/1 Dinosaurs and antigravity: http://www.bearfabrique.org/sauropods/biganims.html \_ it was once "proven" that you would suffocate if you moved faster than 20 miles per hour. What a load of crap. \_ /. \_ It's \. not /.  bill gates \_ Yeah, but read it. It proves that at one time the earth's gravitational pull was much lower. \_ No it doesn't. Read the responses in /. and you'll see refutation of the numbers. \_ Not surprisingly, slashdotters are wrong. Bad math. \_ I can't find anything on http://slashdot.org searching for dinasour. However it seems highly unlikely that they would screw up the trivial math. It is not the math that is problematic but the underlying assumptions. \_ maybe that's because its spelled DINOSAUR! \_ Maybe the atmosphere was richer in oxygen such that muscles could grow much stronger and withstand a much higher force perunitarea than those in modern animals. Maybe the winds blew much stronger than nowadays such that a glider could take off easily. Just some idea.  yuen \_ Wind might help the flyers but doesn't help a 70+ ton beast stand up. I don't know enough anatomy/biology to say anything about your oxygen theory. Where's a bio dynamics major when you need one? 
2000/6/23 [Computer/Theory] UID:18385 Activity:high 
6/1 Prove that gaussian is the only function family closed under convolution. ie, forall f,g in H, exists d in H st d=conv(f,g) ===> H is gaussian family. \_ The difficulty of the problem hinges on the definition of 'function family.' I could, for instance, come up with many 'function families' of one element each for which the above would hold. \_ The difficulty does not hinge on the definition of the function family in the trivial way you suggest; it is deeper. There exist uncountable sets of functions H such that for all f,g in H, f * g is always in H. One example of such a family is the set of sinc functions A sinc (w x) with w in [1 2] and A being any real number. This was what I was trying to point out below. emin \_ Right. The definition could be made too narrow (to make closure easy), or too wide (for instance by making H the set of all continuous functions on R), which would also make closure easy. The point, in both cases, is that we are missing the definition of what a function family is. My guess is that a function family is a set of functions with the same finitely expressible closed form (finitely expressible to avoid the fact that continuous functions on R all have Taylor expansions). \_ Utter nonsense about complete set of real functions deleted. Take a math course, for crying outloud.  motd math censor. \_ fuck you. \_ This isn't true. Let f(x) = g(x) = sinc(x), then f * g = f. Also, if you let f and g be generalized functions then f(x) = delta(x) = g(x) results in f * g = f = g. I belive what you are thinking of is that if you restrict f and g to be probability density functions then Gaussians are the only pdfs closed under convolution. Not sure how to prove this off the top of my head. Please let me know, when you find out. emin \_ Depending on the minutia of the formal definition of PDF, a delta function may be considered a PDF as well (since the integral is defined to be 1....). Then, of course, delta can also be considered a Gaussian with 0 st dev. \_ yeah, sorry i meant pdf. i'm aware of the delta function case, but i guess i'd like to define "in the set of" as "approaches an element in the set H, in some sense". also, i think the proof i have in mind requires L2, which deltas aren't. so my proof was going to look something like: since it's a PDF, it must be the square of some other function. look at the fourier transform of the square of something, and realize that it must be closed under multiplication in the fourier domain. assuming that the function is in L2, that constrains it to a large class of functions, which includes boxes and exponential functions. then i was going to use the fact that this is the fourier of a SQUARE of a something to show that within the exponential function, it must be gaussian. i don't know yet how to get rid of the other families that are closed under multiplication. ali \_ man, do you always have a comment on everything, mr. fricking i know everything about anything?  alum who dislikes you intensely \_ Disliking ability in others is the only form of evil I know.  Ayn Rand #1 fan \_ having ability is one thing, discussing topics like this here is another...go to some math newsgroup, don't clog the motd with this extraneous garbage. \_ Oh bullfuckingshit. Any kind of math discussion is vastly better than most motd discussions, and you know it. It's no crime to be stupid, but it's a shame to be bitter too. \_ The motd has helped me out a lot. Consequently I feel a responsibility to give back and try to help answer questions which I have a clue about. I'm not trying to show off. If you have a suggestion how I can try and contribute to the motd without offending you, please let me know. emin \_ i don't think "alum" was referring to you. you typically provide the kind of insight the rest of the CSUA is completely incapable of providing. ali. \_ i'm talking about ali, not you emin. \_ you stupid shithead. i'm the one who asked the initial question. if you have a problem, feel free to post your own moronic "how do i make text bold using html" questions. ali \_ 'Take your math elsewhere' stupidity deleted, due to its inherent idiocy.  motd math censor. 
2000/5/34 [Computer/Theory] UID:18162 Activity:very high 
4/32 Stat 200a vs Stat 101 vs Stat 134  any comments? Experience with any would be appreciated too. (Assuming that purpose is applications in an arbitrary area of theoretical CS) \sigh, i guess i will wade in: if you let me know what topics you \_ Can you describe the subject matter please? I am too lazy to look it up. are interested in terms of mathspeak [theortical CS to me is more algebra than probability] i might be able to make a recommendation. are you interested in models, time series, probability/measure, game theory or more specific stuff like genetic algorithms, neural nets, baysian stuff/AI? psb \_ bayesian.  cardcarrying bayesian \_ And in general, I hate genetic algorithms. Why are they even in \_ This is why I should have signed my blurbs in motd.  alice the same category as neural nets and bayesian nets? They are completely unprincipled. There are no convergence results for genetic algorithms. In fact, you are not even guaranteed to improve in each iteration. The whole thing is a scruffy atavism be an engineer, i.e., look into EE126 and EE226. from the freeswinging days of AI when any yahoo could get a PhD by writing an Eliza program.  cardcarrying bayesian \_ 134 was kinda fun in a gambler's anonymous kinda way. \_ Real men (and real CS theoreticians) take Stat 205ab. Given \_ and don't forget those strange women that know how to do math. You know, women that are men. \_ "MATH IS HARD! LET'S GO SHOPPING!" Barbie \_ Actually that was 'Malibu Stacy' \_ Real men don't get their panties in a twist over convenient abbreviations like the phrase "real men". \_ Morons. This is why I should have signed my blurbs in motd.  alice \_ Signing it didn't make a difference. Go fuck off. \_ Previous animosity not registered. your choices, however, I'd go with Stat 134, because they seem to talk about more complicated stuff like random processes. It really depends on how you plan to use it. If you just want to know enough to get by in CS 174, then math 55 or CS 70 is more than enough. If you are serious about stat, but your math background is not very good, then do the next best thing and be an engineer, i.e., look into EE126 and EE226. alice \_ Given decent math background (core upperdivs), and completion of cs174 (fwiw). Explain difference between 200 and 205 series; it seemed like 200's are more hardcore than 101/134, and 205 seems less comprehensive from catalog description. \_ 200ab are for business/economy majors, is my understanding. \_ One of the things I noticed is that EE226 and several IEOR classes require the 200's. Biz/econ majors? \_ Ignore the requirements. YOu can pick up whatever 200 stat you need while taking those classes. Like I said, real men take 205ab, but not with Klass next semesterI heard he's a terrible teachertake it with Pitman the year after. You'll run through measure theory in the first hour of the first lecture. (You'll return to it, don't worry.) Then you go on to the good stuff like the Laws and the Processes. It'll be very mathematical, but well worth the effort. This is all hearsay, btw, I haven't yet taken 205. To the best of my knowledge, you don't need such a solid foundation in stat for any CS theory courses. However, you'll probably have a much more transcendental experience in Sinclair's MCMC class with some 205 under your belt. Mmmm, the Sinclair Eclair.  alice \_ For fun with AI, take Michael Jordan's year long series of courses on graphical models. They are crosslisted as graduate stat and graduate CS. Very cool stuff.  ilyas \_ I thought he was a retired basketball star and extalk show host? \_ No, he is actually a star AI researcher, freshly stolen from MIT. He works on variational approximation methods that allow one to deal with very large graphical models.  ilyas \_ Oh yeah. Forgot about Mike's classes. They will be excellent. Knowing him, he's probably going to go through calculus of variations, variational methods, mean field theory, and general statistical mechanics in one lecture. It'll be a blast.  alice \_ My ass still hurts from last time.  ilyas \_ Please don't make me picture your ass, ilyas. 
2000/4/1213 [Academia/GradSchool, Computer/Theory] UID:17980 Activity:very high 
4/11 emarkp, ilyas, ali, are you guys in grad skool? \_ tell us about the stars ilyas! \_ I am not in grad school because I lack the grades and references to make it to a first tier school. Thus, I am trying to get my foot in the door by doing something reasonably impressive in the industry.  ilyas \_ the stars!  Irami Osei Frimmpong \_ grad schools don't look so favorably upon most industry experience. They want to see letters from people they know in industrial research or academic research (or so says Aiken...) Besides, ilya, I always figured you more as the capitalist worker exploitation type... hmm... ;) brg \_ Industrial research is, unless you work at IBM, an oxymoron. \_ that's bullshit: at&t bell labs, hp labs, microsoft, and of course the many biotech firms that have been founded by research scientists with the intent of doing research in a specific direction. i won't even mention all the major chemical companies with huge research divisions, such as Dow. \_ Okay, Microsoft research is an oxymoron. \_ no, it's not. they have hiret some badass quantum computing theorists. if you .ook at the paters in phys rev a on quantum information, alot of them come from microsoft peopl. \_ Quantum computing research is an oxymoron. And learn some damn grammar/spelling. \_ Microsoft employs some of the world's best AI researchers like David Heckerman, and Christopher Bishop.  ilyas \_ Is that why those dancing paper clips seem so intellegent? \_ Aiken is right. \_ In that case I lose.  ilyas \_ the industry is full of $$$!!! why the #*( would you go back 2 skool??? \_ define first tier and second tier (first tier == top 5? second tier == the next 10 after that?) \_ first tier == { "Berkeley", "Stanford", "Carnegie Mellon", "MIT" }; // No particular order second tier == { "Cornell", // VERY close to first tier "Brown", "Princeton", ... (there's quite a few and even if I tried to list them all, I'd miss one) }; dans \_ i'd say you should put Georgia Tech in there if you're going to put CMU. the top 3, everyone says is Stanford, Berkeley MIT. \_ Georgia Tech doesn't come close to CMU or even Cornell. And in reality, Stanford is considered top tier even though schools like Georgia Tech, Utah, and UNC have arguably better programs... \_ It must be listed in alphabetical order! Berkeley, CMU, MIT, Stanfraud.  yuen \_ What about CalTech? \_ why are you asking these people specifically? and you should sign your name if you want these people to respond. \_ I don't know why you're asking about me specifically. But no, I'm in industry with a BA in CS. emarkp \_ well, you sound like a nweaverwannabe and he's a grad stn \_ What does that even mean? Do you mean that not using ad hominem attacks makes me look like a grad student? Small world you live in isn't it? (And sign your name.) emarkp 
2000/4/3 [Science/Electric, Computer/Theory] UID:17915 Activity:nil 
4/3 Why doesn't CAL and/or CSUA participate in Robot Wars? \_ Define Robot Wars. IEEE supports MicroMouse. Don't know much about it myself but it might be along the same lines. 
2000/3/29 [Reference/History/WW2/Germany, Computer/Theory] UID:17882 Activity:nil 
3/27 fibonacci sucks big dick \_ I wish. but mine wont fit \_ You wish a guy suck your dick? \_ You writing your grammar wish good? \_ Why not? Who cares who does it? Lips & tounge are lips & tounge. \_ Sigh. \_ Ahhh. \_ formatting hard so Is? I all I'll start purging think contentfree that comments formatted poorly are. formatting nazi \_ Do it. Blow 'em away! \_ what the fuck is a tounge? \_ An extremely clever anagram. pld 
2000/3/29 [Computer/Theory] UID:17881 Activity:nil 
3/28 I nuke that whole fibonacci discussion. I think it was definitely created to torture students throughout junior high, high school, and college. And that guy who called me stupid...kiss my ass bitch! Shit...if it was so *easy* and so *simple* why was there a big ass post here on the motd?? Fuck fibonacci, go talk about the Fourier Series, Laplace transforms, Legendre polynomials, etc, etc, just no more fib...it grows tiresome. Or talk about Reality brand female condoms...better or worse than the norm? blah blah blah... \_ i think you nuked it because people were exposing your stupidity too much. 
2000/3/2729 [Computer/Theory] UID:17867 Activity:insanely high 
3/27 What's the running time (bigO) of fibonacci series? I can't find it on CLR. It only talks about the golden ratio shit. \_ F_n=(p^n(1p)^n)/sqrt(5), p=the golden ratio. Galen \_ CLR was the most worthless piece of shit book I have spent my money on. It's worth more as fire wood or a shooting target than an academic textbood. as an academic textbook. \_ Go read some Sesame Street literature, then. There's a good REASON that CLR is the single most popular intro algorithms text in the world. \_ O(1). \_ dumbass \_ A numeric sequence can't have a running time, only an algorithm can. Please restate your question. Also, Fibonacci numbers are a sequence, not a series. \_ How about "What's the running time of evaluating the n+1th number in the Fibonacci sequence F(n) by using the recursive algorith F(n)=F(n1)+F(n2), F(1)=F(0)=1?" \_ The naive implementation is exponential, the slightly less naive memoizing implementation is linear. \_ The cool implementation is constant time. You can derive a formula like F(n) = a * b^n + c * d^n. I don't remember the coefficients, but you can look them up or derive them. If you need the formula and you can't find it, email me and I'll take the time to derive the constants. emin \_ F_n=(p^n+(1p)^n)/sqrt(5), p=the golden ratio. Galen \_ uh, if you have to use "to the power of n", for arbitrary n, i don't think it's constant time...nice try though. \_ "To the power of" is O(1). Nice try though. \_ I'm impressed...with your stupidity. I mean, you can't really be THAT clueless can you? You'd have to be TRYING to be stupid to come off that well. I really hope you were expressing something in an odd and painfully obtuse way rather than blithering utter and complete nonsense. \_ Go read a book, moron. Let's see you do O(1) ANYTHING. O(1) means you don't even have time to read the entire input. Exponentiation in arbitrary precision is O(log) with respect to the exponent. Actually something like O(n^3) w.r.t. actual input size, since the base has to be multiplied by itself O(log(exponent)) times. \_ BTW, what's the Fibonacci series famous for? It might be useful for teaching computer science, but for math itself what's it useful for? \_ It comes up in the wierdest places. People have used the fibonacci sequence to analyze the stock market (successfully). \_ yep. there is a journal called the "fiboonacci quarterly" devoted entierely to such weird cases. it should be in the math library if you're interested. \_ If the stock market was that easily analysed, we'd all be billionaires. \_ you mean you're *not* a billionaire? \_ I meant the rest of you peasants. \_ Modeling the reproduction rate of rabbits. \_ yeah, i hear you on that one. the most i've ever seen it for is the continuous torture of students throught junior high, high school, and college. if anybody can actually point out a url which proves that this thing actually has a practical application in reality, i'd be really pleased to see it... die, fibonacci, die!!! \_ Torture? Fib series is trivial. Where'd you go to school? Dummy High? \_ C is for Coo^WYour Grade and it's good enough for you! \_ If that's the most you've ever seen it used for, you must be an inexperienced college kid with little view of the wider world beyond the school walls. That puts you in the right place. \_ for modeling the rate of spawning of new rabits in the wild ali \_ This is important because....? \_ because it also models the rate at which i and my progeny have been doing your mom for the past two years. ali \_ Try Fibonacci heaps. \_ Whatever. The complexity of the heap operations incur so much overhead that they eventually end up costing a lot more to implement than with an alternative and simpler implementation thus cancelling out any improvements gained from lower running times. This is as useless as Straussman's (sorry if I spelled his name wrong Ilyas). Big deal. We can matrix multiply in O(n^2.7) time instead of O(n^3). It turns out to run slower on most machines anyways so why bother. \_ what the fuck are you talking about? name another data structure that implements priority queues with logn insert and 1 retrive. the "simpler" alternative you have in mind, is it an insertion sort or something stupid? \_ Maybe he is talking about standard heaps which have logn insert. \_ Strassen's is getting to be O(n^2.3) now. And even the O(n^2.7) implementation becomes faster than normal multiplication for sufficiently large arrays on the machines where 170 homework is done, at least.  ilyas \_ The best known, n^2.376, isn't based on StrassenWinograd original... not that it matters, of course... \_ Do you have any idea how big the matrices for interesting realworld problems tend to be? Do you have any idea how much less (n^2.7) is than (n^3)? blojo \_ Why, do YOU? Because it isn't all that large. Even a lame implementation passes the threshold at about 150x150 matrices. A _lot_ of problems involve much larger matrices which are sufficiently dense for sparse methods not to work. \_ Are you arguing for or against Strassen's? \_ That's not how it is in my industry (games). Nobody doing a complex system will write bruteforce matrix math and expect it to run okay. It's all about multigrid methods, frequencyspace operations, you name it. blojo \_ while (player_still_alive()) {draw_dot(x,y);} \_ Ideas similiar to Strassen's algorithm are used for fast FIR filtering which is widely used in all sorts of DSP algorithms. I don't know if people thought of fast FIR filtering first or Strassen's algoirthm first. However, the point is that the ideas behind Strassen's algorithm turn out to be very useful. IMHO the reason to study algorithms (even weird stuff like Strassen's algorithm and Fibonacci heaps) is so that when you have a particular practical problem you can come up with clever ideas. emin \_ Strassen was there first. \_ I'm sorry, Strassen's should *never* be faster on modern architectures, where multiplies are as fast as adds. Strassen's trades off more adds for fewer multiplies. As long as you've blocked your implementation for cache and registers, and copied when you have lowassociativity caches or TLB, then you're OK without Strassen's. nick \_ Uh, go read the book again. Strassen optimizes away MATRIX multiplies, NOT scalar multiplies. If we knew how to multiply matrices as fast as we can add them (O(n^2)), all of Strassen's instead of 8 muls, you have 7 muls. ali stuff would be obsolete, yes.... But we _don't_ \_ Isn't this one of the advantages of a Quantum Computer over standard computers. \_ Oh, shut the fuck up, and consult the literature. \_ Find a new book. At the instruction level, a multiply is a multiply. Thank you for playing. \_ Wow. Are you just a total idiot or what? For a square matrix of size n x n, addition or scalar multiplication requires O(n^2) operations, whereas naive matrix multiplication requires O(n^3) opations. Yes, add and mul are single instructions, but Strassen attempts to reduce the number of those instructions. \_ This was exactly my point. So there's no reason, as you said, why we should use add instead of mult. \_ because while add and mult are equally fast, add is faster than mult. Duh! pld \_ look, strassen's replaces a multiply by 14 adds so that instead of 8 matrix muls, you have 7 matrix muls. ali \_ Thank you for reaffirming my point that an mult is the same difficulty computationally as an add, so there's no need for all these Strassen methods and their variants. \_ it replaces 1 matrix multiply by 14 matrix adds. A matrix mutliply (brute force) is O(n^3). A matrix add is O(n^2). are you really this dense? \_ But they're the same if n is 1, so fuck you! \_ I think you can use a generating function z() to compute the nth fib. number in constant time hence the O(1). \_ Formatting fixed. \_ The above function IS the generating function and it's NOT O(1) with respect to input size. Read above posts. \_ Think outside the box. The answer is there if only you would reach for it. \_ i think bigdickbastardpedant's complaint is that you can't generally compute e^x in constant time, because you need to compute at least a taylor series or something, and the number of terms used does depend on x, as larger x's are more poorly approximated with few terms. i don't know of a way to compute e^x in time independent of x either. ali \_ "Hello, Pedant." *clickety click* \_ Well The Pedant needs to go reread exactly what O() notation is all about and how to properly compute it. The size of the input is unrelated to the O(), therefore The Pedant is not only a pedantic idiot, but is simply flat out wrong as well. \_ you must have this strange verison of O() notation that depends on the phase of the moon or something. \_ If sorting is O(n log n), what does n measure? lumen? pld \_ fibonacci sucks big dick \_ I wish. but mine wont fit \_ You wish a guy suck your dick? \_ You wish your writing good grammar? \_ Why not? Who cares who does it? Lips & tounge are lips & tounge. \_ Sigh. \_ Ahhh. \_ Is formatting so hard? I think I'll start purging all contentfree comments that are poorly formatted. formatting nazi \_ Do it. Blow 'em away! \_ what the fuck is a tounge? \_ An extremely clever anagram. 
2000/3/67 [Computer, Computer/Theory] UID:17703 Activity:very high 
3/6 Computer science theory shmatheory. I'm not a phd candidate like nweaver or dpetrou but I'm making #)(*$)# more $$$ than they are. \_ You know what they say about fools and their money... brg \_ nweaver probably still has far more $$$ than you (gotta love rich parents) and dpetrou is definitely far more intelligent and happier with his life than you are. All you have is money, and that is fleeting. \_ Nice to know nwaever can afford to mooch off his parents and join the intelligencia and look down upon us poor slobs who have to *work* for our money. You have no basis of comparison between dpetrou and the anonymous person above, so nothing further need be said on that. not the above \ well. love is still better in the end than $$. ( future IPO millionaire) \_ money can't buy happiness, but it can buy... legos! yesh! current IPO millionaire \_ Money can't buy love but it can buy the alleviation of most forms of personal suffering. After your basics are met and are guaranteed to remain so for life, any misery remaining is your own fault. Go see a therapist. \_ Money can't buy yer mom, because she is free. 
2000/3/12 [Computer/Theory] UID:17666 Activity:high 
3/1 Today we'll be talking about artificial intelligence by useing examples from The Matrix with the help of Stuart Russell. 308 LeConte, 5pm. all are welcome to attend! sofia \_ Someone go there and protest that Matrix demonstrates nothing about AI. Watch 2001 instead. \_ One more reason to STAY OUT OF INDUSTRY, COME BACK TO ACADEMIA!!! \_ One more reason to STAY OUT OF SINDUSTRY, COME BACK TO ACADEMIA!!! \_ Does academia include free use of a spell checker? 
2000/3/12 [Computer/SW/Languages/Java, Computer/Theory] UID:17663 Activity:high 
2/29 What's the difference between a contextfreegrammar and a contextsensitivegrammar? \_ Contextfree: a rule maps a nonterminal onto a string of nonterminals and terminals; decidable by a pushdown automaton Contextsensitive: a rule maps a string of terminals and nonterminals to a _longer_ string of terminals and nonterminals; decidable by a linearbounded turing machine, but not by a PDA. Universal: a rule maps a string of terminals and nonterminals onto any other string of terminals and nonterminals (aka Universal Rewrite Rules), decidable by a TM, but not always by a LBTM/PDA. For more, RTFM; the motd is not a math book. alexf \_ I wonder who is asking 172 questions on the motd? brg \_ A lazy idiot.  ilyas \_ Yes, and anyone dumb enough to take 172 is just an idiot. \_ Hey, for the record, 172 was my favourite class at Cal. But asking such ... classrelated questions on the motd seems pretty lame to me. The least you could do is troll well.  ilyas \_ And mie favourite klass was Collej Righting 1A becouse we lerned to spel words like favourite. \_ Anyone who doesn't like the same classes as ilyas is a troll. \_ Anyone who says stupid shit like: '\_ Yes, and anyone dumb enough to take 172 is just an idiot.' is a troll.  ilyas \_ ilyas hath spoken and so shall it be! \_ I liked 172 a lot, even though I got my ass kicked and I am definetely not into math... muchandr \_ more like 164 I bet. \_ 164 almost never covers CSLs/CSGs \) why did everyone bitch about my 164 class then? aspo \_ Who cares? \_ 164 is boring. 264 is more interesting. 
2000/2/2829 [Computer/SW/Languages/Java, Computer/SW/Languages/Perl, Computer/Theory] UID:17642 Activity:insanely high 
2/28 If the pushdown automton can accept the language consisting of all words in the form (a^n b^n, n=1,2,3,...), how come no pushdown automaton, deterministic or nondeterministic can accept (a^n b^n c^n, n=1,2,3,...)? How would a linear bounded automaton accept this language? \_ Just write a perl script for some computer. Make k large enough \_ Do you own homework. to make 3*n*k be equal to the RAM size of the computer. \_ 'your' \_ 'typo' \_ 'idiot' \_ Think of this the next time you typo, idiot. "That's how." \_ and people wonder why the csua is a sysadmin farm \_ pumping lemma pumping lemma pumping lemma, hehe. \_ Do your own homework. \_ Are you saying no one has shown that it can, or are you saying someone has formally proved that it can't? \_ "Someone" has formally proved that it can't, and YOU CAN, TOO! With just a small 4unit investment into an intro complexity class, YOU TOO can go around saying cool words like "pumping lemma", "LBTM", "Savitch", and "EXPfuckingSPACE." Take 172, and CHANGE YOUR LIFE FOR THE BETTER! alexf (real answer: pump) \_ I did take 172 in Spring '91. I just had a lousy professor (Anvari). \_ How relevant is complexity, architecture, automata&languates, logic&systems design, theory of computation? NONE. I took a c++/Java class and now I'm making 80K/year. not CS major \_ relevant to what? just because most people in computers dont need to understand fundamentals doesn't make it important. this is like all those idiots who think that because they make tons of money in software hardeware is useless. without people to concieve of new kinds of computing devices who understa nd theory and people to build those devices, you would not have anything to work on for your 80k/yr. no one is going to pay you 80k/yr to write addition algorithms on Babbage's difference engine. not relevant to *your*life != not relevant. \_ I happily make tons of money in a nonhardware field and fully appreciate all the hardware toys that come my way. Not a CS major and making more than the clown above who doesn't get it. Every advance in either software or hardware is good for everyone's bottomline. \_ Is your code maintainable and extensible? Probably not... \_ Are you in school and making +80K? Probably not. Are you in industry and making +80K? Probably. \_ Industry is for brainless people and academia is for brainy people. Industry sucks and academia rules. B1FF G0 BAC< 2 SK00L!!!!!!! \_ "I work for academia and get paid less to do less, therefore I am morally superior and more intelligent than my industry counterpart." This is such a total crock. It was hard to get a good first industry job coming from academia because in industry they know the book worms don't know how to produce, can't work and can't get along well with others. Academia isnt even academia anymore. How many rooms and buildings are named after companies or CEOish sponsors? How many projects on compas are paid for in whole or in part by industry? You think that's free money from the goodness of USA INC's heart? Academia is the cheap research arm of industry, today. Traditional academics are dead. Long dead. \_ Idiot. It's the long term research in the industry that is dead. Go away, you superficial software drone. 
2000/2/2526 [Computer/Theory] UID:17618 Activity:low 
2/25 Next week's topic in the science fiction DeCal is Artificial Intelligence. We'll be watching The Matrix on Monday @6pm 160 Dwinelle, and Stuart Russell will lead discussion on Wednessday @5pm 308 LeConte. All are welcome to attend either class. sofia 
2000/2/1617 [Computer/Theory] UID:17523 Activity:insanely high 
2/16 I always ask this simple computer science question: What is the running time of an insertion sort? 50% of the time, the candidates can't even answer that question correctly, yet they are making 5060K/year. What is wrong with this industry? \_ Maybe you're the one with the wrong answer. \_ Labour shortage. I used to be proud of how much I was making until I realized a lot of idiots make even more. \_ What IS the running time of an insertion sort? \_ O(n^2) \_ Wrong answer. Right answer: "Under what performance model?" nick \_ Average case? \_ That's just the issue of which asymptote to use, not which performance model to use. nick \_ Enlighten uswhat do you mean by "performance model"? \_ RAM, PRAM, PMH, LOGP, BSP. When is merge sort likely to be faster than quicksort? It depends on which model you use. RAM says quicksort, PMH says merge sort on machines with caches. (independent of the partially sorted issue). \_ I guess I'm just an idiot. What do the acronyms mean? \_ You're fired. The answer was O(n) because the data was presorted and inserting one more element in a presorted list isn't O(n^2). More specifically, the exact correct answer is that not enough information has been provided to correctly answer the question. an idiot and is asking the wrong question. I *HATE* having a \_ You're fired if you can't figure out how to insert into a sorted list in less than O(n). right answer and ask it properly. This goes out to all dumb fucks out there who don't know what they're asking. You're not penalising people for lack of clue. You're penalising them for not being able to read your mind! That's why so many job reqs. \_ You need a 'special' list for that. Lists don't have random access, and array don't have constant time simple question across without wrongly and unfairly assuming it insertion, so you have to marry them. choosing the wrong interpretation of the many valid possible from your _very_ _poorly_ designed question. How did you graduate from Cal and get into a position to interview when you're not capable of asking an intelligent question in a way that can be understood by others in the field? Sheesh. I really hate the arrogant inarticulate geeks who are incapable of seeing any fault in themself. The fragile ego can't take it. \_ Did you know that Phillip Nunez invented the skiplist? \_ By the way, what is the running time of inserting n elements into a vector?  ilyas \_ tjbP? The person above "ragging" on people getting the wrong answer is an IDIOT and is asking the wrong "question". I *HATE* having a totally *STUPID* interviewer like that. If you're going to ask any sort of technical "questions", you better fucking *know* the right "answer" and ask it _properly_. This GOES OUT to all dumb fucks out there who don't know what they're ASKING. You're not penalising people for lack of CLUE. You're penalising them for not being able to "read" your MIND! That's why so many job reqs. specifically say "excellent written and _oral_ communications skills". That's to avoid _dummy heads_ like you who can't get a simple "question" across without wrongly and unfairly ASSUMING it \_ When I grow up, I want to go to Bovine University! is the other person's fault for _not_ reading your mind and choosing the wrong INTERPRETATION of the many valid possible from your _very_ _poorly_ designed question. How did you GRADUATE from "Cal" and get into a position to INTERVIEW when you're not _capable_ of asking an intelligent "question" in a way that can be UNDERSTOOD by others in the field? _Sheesh_. I really hate the arrogant inarticulate "geeks" who are incapable of seeing any fault in _themself_. The fragile ego CAN'T take it. \_ Right on. \_ Exactly. Besides, known questions can be looked up. Better questions investigate how a candidate thinks and how he/she can communicate with you. \_ The appendix to the "Deep C Secrets" book has a good summary of answers to some of the most common MS & SV programmer interview questions. \_ More importantly, who cares? sysadmin over 90k \_ 5060k? 65k is starting salary nowadays... \_ shit, i got screwed again \_ A while ago I interviewed a few people for software engineering jobs. One of the easy questions I asked was: What is the running time of quicksort? Four out of five candidates got the right answer. The fifth had a certificate in computer science from Microsoft University. He didn't have a clue about \_ When I grow up, I'm going to Bovine University! quicksort. Did you notice a corellation between university degrees and people knowing about insertion sort? emin \_ What is your answer? \_ The answer I was looking for would be O(n log(n)). Answers like O(n log(n)) in average case and O(n^2) in the worst case would be even better. A discussion of computing models would have been fine as well. Basically I wanted to make sure we hired people who had a clue about efficiency, not just people who knew the syntax of a programming language without knowing anything about efficiency. It was not a trick question. It was an easy question. emin \_ Actually, if that's the answer you're looking for, you better refine your question. Instead of asking for the running time, which is subject to interpretation (computing models, average case vs. worst case, etc.), perhaps you should ask someone to actually perform an analysis of the running time for a given algorithm. dans \_ I agree. I also asked algorithm analysis questions. My point I've known people who interview for a programmer job without any clue of algorithms. You ask them about quicksort and they stare at you blankly. \_ I wasnt staring at *you*. 
2000/2/5 [Computer/SW/Languages, Computer/Theory] UID:17432 Activity:nil 
2/3 /tmp/tjbresume.pdf \_ Isn't anyone going to do anything?? Scratch one up for the bad guys...Somebody should anonymously be slipping a transcript of that tirade of his under the door of the math department. Why is it the world that assholes like this always win?  angry anon \_ Actually, we should find out how danh got the resume and see if he was breaking any laws in posting it. It looks like a scan or fax. Which means he probably had a physical copy. He also could have had at least the consideration to destroy the home address and phone number from the posted file. That alone could be considered illegal. \_ When you make a resume and submit it, you're doing so under the assumption that an entire organization of people and possibly multiple (contracting firm and the place they send you) are going to look at it, enter it in mass databases, etc. You _want_ lots of people to see your resume. It isn't confidential information. \_ Thanks for overwriting my comments. Yes you expect people who may be considering hiring you to read the resume. You do *not* in general expect to see it posted to a generalaccess website. I wouldn't put my home phone, cell phone and address on a website, but I would put an email address. I think tjb could take danh to court. \_ blah blah blah. its the motd. no one intentionally overwrote your whine. tjb can't prove any harm and courts are expensive. once you release your resume you have every expectation that it _will_ become public info. youre a fool otherwise. tough shit. \_ Please show me where I stated that the overwriting was intentional. Or, what, only you can complain about incompetence? It's disturbing that you equate illegal with tjb's abilities to pay for court proceedings. Just because tjb was a jerk doesn't mean that gives anyone license to stalk him. \_ whine whine blah blah whine whine. It's the motd. things get overwritten. whine blah blah, bleah. Hardly worth mentioning no matter how it happened. whine blah whine. \_ Posting a resume isn't stalking. Name one law actually broken. If sharing information about people was illegal, credit bureaus couldn't exist or make so much money selling your reports to credit card companies. \_ But posting the home address/phone number along with deprecating remarks and suggestions to harass tjb (all of that here on the motd) is stalking. \_ Under what imaginary law? \_ So tjb can call the cops if he wants. Good luck with that. No it isn't stalking. At worst it shows bad taste on dan's part. If he really feels like doing something, he should go get some student conduct board on dan. Again, good luck. /tmp on a limited access computer is hardly public, anyway. Can you say, "Over sensitive", "Mountain/molehile", or "Wah! Mommy!"? \_ Look again you moron. The file in /tmp is a symlink to danh's webpage. That's most definitely public. \_ Yeah, dan's soda web pages are getting millions of hits a day. Like I said, waah waah waah. Go bitch to someone who might care, like your mother. \_ epsample is no more \_ since when is it illegal to scan in someone's resume? maybe i want to hire him. I think tjb qualifies \_ You needed a new asshole? \_ How can you have GSA appointment when youre an undergrad? by definition he has to be an *undergraduate* teaching assistant. Unless the math department is now handing out graduate appointments to sophmores... \_ Happens all the time. Occasionally they call them "UGSIs" but they do everything the same except they don't get a tuition credit (at Cal) \_ d00d, m4d B4S1C sK1LLz!!1!! \_ Is this just me or not? I just looked at it. IMHO, I am much much more impressed by most of the resumes of various people posted on <DEAD>soda.csua/~membername<DEAD> tjb's resume looks like an average nerd type resume, the kind that usually go to grad school after graduating. I didn't see much accomplishments (aside from just getting good dull grades) in there. And those grades aren't the greatest ones I have seen before. \_ Anyone want to run a check on the 'first person in history of university' to TA math as an undergrad claim? I'm having a rather hard time believing that. \_ maybe he's right about the math 55 claim, since i bet most serious math students don't bother taking math 55 unless they are CS majors \_ my friend TA'd math 16a,b as an undergrad, and had all the same responsibilities as a normal GSI. it is quite common. the math department has somewhat of a labor shortage and honestly does not give a shit about undergraduate education(yes, i have evidence to back this up, and no, i'm not saying there aren't some great individual teachers in the department.) the pay sucks, the workload is high, and they'll hire anyone who knows calculus and will put up with the bullshit. mathmajor alumnus \_ it's not supposed to impress you. there are several strange things, for instance, he made double the $$$ delivering asian food to the suburbs (??) than he did programming computers. plus he has MAD BREAKIN' SKILLS. I haven't seen anyone breakdance since I rented Krush Groove 10 years ago. I would love to know under what name he practices his turntablism (for lack of a less stupider sounding name) skills, DJ TJB just doesn't fill me with awe. Trevor are you reading this? csua member posting anonymously, just to avoid moronic flames from this asshole \_ what's so wrong with the type of person who goes to grad school after graduating? \_ May be he meant that academic success is not all that matters in one's life or even on the resumes. \_ What's worse? A moron? or the moron who goes thru the trouble of posting a moron resume and posting to the motd about it? Okay, so danh is not a moron, but still, makes ya wonder about these CSUAers  do they "got life"? \_ i am a moron today, but I had a lot of fun figuring out how to make a pdf finally  danh \_ Making a pdf was fun? \_ Since it's scanned, wouldn't a jpg or even gif have been smaller? \_ In framemaker, it is a simple "save as" routine. \_ shut the fuck up, cmlee. tom \_ obsessing over small details is not a moron trait; it is a geek trait. Bragging on your resume about your mad breakin' skills or whatever and listing a food delivery job as work experience is a moron trait. And fucking with people who have a presence in every corner of the industry you someday plan to work in is DEFINITELY a moron trait. illuminatus \_ This is the part I like. All tech resumes pass over my desk before anyone else sees them. Only the ones that don't go in my trash can have a chance. This includes people referred by current employees (as if anyone would refer this guy). I post the worst ones outside my office for everyone to laugh at instead of the typical array of Dilbert comics. patiently waiting for tjb's resume to post on door \_ what company do you work for? it's assholes like you, not tjb, who make people leave computers , prefering a pay cut to totally defective pricks as coworkers. scientist making 1/5 your salary who doesnt work with pricks \_ I'm an asshole for what exactly? Posting the funny resumes or throwing out the 100s of bad ones I get every week? whine blah bleah whatever. I doubt you make enough to make 1/5th my salary. But I'll pay you $5 for each car you properly wash. That should easily double your salary. But don't come near me. I don't have time to waste on ivory tower whiners like you. Just wash the vehicles and go away. \_ I agree with the scientist that you are an anal sphincter, and lacks basic social courtesy. Not only are you an asshole, by attacking tjb, you are also a hypocrite since you are no better than him. \_ Still waiting to hear exactly what my crime(s) are/is. Until then, you're easily dismissed as the whiney spewing potty mouthed child you seem to be. 
2000/1/2426 [Computer/Theory] UID:17314 Activity:high 
1/24 Should I expect to put more work in cs170 or m113? \_ 113 was cake. One of the easier upper div classes I took. 170 had piles of work but it wasn't hard. It only takes time and lithium to get through any upper div math class. \_ lithium? what is this a reference to? \_ Also a great Nirvana song! Woo hoo! \_ To the bulk quantities of drugs shipped to the math dept daily to keep the profs semistable. Lithium is for people with various mental stability problems usually lumped into the meaningless title of "schizophrenic" (which is totally different from multiple personality disorder). \_ Ummm, no. Lithium is used in treating bipolar disorder, aka manicdepression. It's a mood stabilizer. Often, antipsychotics (Haldol as an example) are used to treat schizophrenia. And the two disorders are completely different things. I don't know if I'd call them meaningless, though, real folks suffer from these things. bipolar sodan \_ I'm glad you know a single use for lithium. I would correct you on the details but wouldn't want to upset your delicate condition. "Meaningless" in terms of "a diagnosis of schizophrenia" doesn't _mean_ anything. Psychiatrists tend to call any and everything they can't properly diagnose some form of schizophrenia. Thus the term is meaningless not the condition itself. \_ actually schizophrenia is very specific, it just happens to have many different effects. \_ "Yeah it's like totally specific, but has like 851+ effects and they like know all about it which is why anything they can't properly diagnose is called schizophrenia!! YEAH!" Uhm, no. You're simply wrong. Think about it. \_ Don't inbreds have an excess of lithium in their bloodstreams? John \_ I'll have to ask my cousin/brother/uncle. \_ M113, by far \_ The CS geeks I know would say 113 because it isn't familiar to them, the math geeks I know would say 113 because "CS is so watered down and math is for real" except these same math geeks always talk about how incredibly easy 113 is compared to their other classes. Ask ilyas about his mathematical purity of essence and the lack of similar purity on the part of mere CS losers (in his mind). \_ I never took either class while at school, but when I started working, I ended reading most of the CS 170 book because it had so much useful and interesting stuff. I don't know which is more work, but if you are in CS you'll probably get more out of cs170 than math113. By the way, I've always meant to teach myself abstract algebra, but never quite got around to it. Can someone point out some incentives for classes like m113? \_ Teaches you to hold the same thing in your mind for longer than the default american attention span of 3 seconds.  ilyas \_ Damn ilyas, I never thought YOU'D stoop to such an idiotic troll. What a bummer. :( \_ You know I was reading this newspaper article about new, 'tougher' standardised tests in california. This article was talking about how some students 'cried in the middle of the test and gave up' because they thought the test was too hard. Call the whole thing a troll if you will but the american public school system really doesn't teach things like attention span, or proper english, or foundations of math, or whatever very well. This is even more surprising because american colleges are in general excellent and emulated by the rest of the world.  ilyas \_ Don't confuse the California precollege system with the rest of the country. CA doesn't have a school system. It has a 12 year baby sitting service. ilyas, you're either trolling or don't know what you're talking about or both. The anecdotal "some student in this one article" isn't proof of anything. I thought you were some sort of self proclaimed super rational logic genius? Stop trolling. American with longer attention span than ilyas \_ Is the public school system any good in other states? I would be very interested in any information on this. And stop trolling yourself, rationality is intractable.  ilyas \_ Take a look at grad theory class prereqs. Quite a few ask for 113. \_ Traditional uses for 113cryptography, quantum physics, graph theory (networks). Pretty useful stuff, even if you never do the quantum. \_ Quantum physics and graph theory my ass. Cryptography, yes. But never in the class did they mention anything remotely close to quantum physics or graphs. \_ Moron. Did your 1st grade arithmetic teacher happen to go over all uses of arithmetic for, say, quantum physics? You're not going to deny that it's used, in quantum physics and elsewhere, right? Just because most 113 profs don't bring up those subjects, doesn't mean it's not directly applicable to them. It is. Ask anyone educated enough in either field. \_ ok fancy pants. if you know so much, tell me exactly what part of 113(that isn't covered in 110) is needed in anyway for quantum mechanics besides crystalography wich i dont count. ive taken 113 and physics137a/b, and i dont see any connection . youre lame. you human paraquat. \_ D00D!!1! U R SO K3WL!!! 2 S3M3S+3RZ OF QUANTUM M3CH!!1!! KRAD!11!!!!! Read a book: Mirman, "Group Theoretical Foundations of Quantum Mech", ISBN 1560722487 Come back when you're done reading. \_ I find this offensive.  nonhuman paraquat \_ Math 113 is less useful and is an easy class. So, I'd put more work into CS 170. dim (Applied Math) \_ I think they're about equal. Put more work into whichever you find interesting. If you're in CS, you'll probably pick up lots of 170 by osmosis later on, so I'd pay more attention to 113 pld \ in my vast experience at the upper div and grad level, math classes vary greatly in terms of difficulty of grading and work load by professor ... that difference swamps any "inherent" differences between various classes. psb \_ He's back! psb #1 Fan \_ Maybe. \_ Perhaps. \_ Yes. \_ Psb. \_ Qed. \_ ED! ED! ED is the STANDARD! Almostthreeletter abbreviation. \_ ED is not an abbreviation. \_ EDitor? Hello? Are you dumb? \_ Huh? I don't get it!!!! 
2000/1/21 [Computer/Theory] UID:17287 Activity:high 
1/21 Just curious, is computer science considered pure science (as in medical science research) or is it considered engineering (as in Electrical Engineering)? \_ Most classes are a mix, though in different quantites. The ABET classifications give a rough approximation. CS 170 is one of the last classes in the world to be considred engineering. Nor is CS 188. While classes like CS 150 and 152 are heavily engineering oriented. \_ I guess these days if you concentrate on theory stuff like the 170 series or AI, it is science. Otherwise it's more like engineering. \_ AI = applied math complexity = applied math \_ no, math = applied complexity compilers = applied math architcture = applied math 
2000/1/21 [Computer/Theory, Academia/Berkeley/CSUA/Motd] UID:17283 Activity:insanely high 
1/20 MOTD Poll: Trevor Buckingham has been made a TA in Math 55 this semester. IF you have been following the prior developments, then do you think that: a. Math Dept should be notified of his prior history by any individual willing to do so. b. Math Dept should be notified of his prior history by the CSUA as the representative body of potentially affected students c. Math Dept should not be notified of his prior history. This is a serious issue; please withhold ballot stuffing, etc. Results will be considered final at noon on Saturday. a: 4.5 + i b: 2 c: 2.5  i \_ with a minimum of swearing and ranting, who is this guy? alum \_ The CSUA ought not to get involved, and moreover, if TB can teach he should  being subservient to the teaching establishment is hardly a prerequisite for being a good teacher. It's well known there is a shortage of even minimally competent TAs. And if he does anything too grossly out of line, he can be fired; for those who are unsympathetic, you might say that he now has a contractual obligation to put his money where his mouth is. brg \_ regarding the half votes: anyone "should" be able to do whatever the heck they want, whether it's reporting TB or not, but a notice from the CSUA smacks of ASUC Senate opinions, and you really don't want anything to do with that do you? \_ The CSUA has no business getting involved and can only get in trouble doing so. Just print out paolo's web page & last semesters posts and slip them under the prof's door. \_ How about letting him demonstrate whether or not he can teach? \_ This strikes me as a cowardly way to do things. pld \_ Wow. I just read that file. Someday, I hope to have the opportunity to reject him as a job applicant. I wonder if he realizes how many people on soda are in such positions in industry? \_ Go to work in academia (okay, I'm channeling kchang here) and deny soda motd hosers like this guy the chance to reject you as a job applicant. \_ He got an A+ in Kahan's class. That counts for a lot. \_ He just seems to be immature. I don't think that necessarily makes him a bad TA.  mikeym \_ This is a good opportunity for some smartass kid to give him a taste of his own medicine. \_ That would be great, except that given his stunning ability to consider other people's points of view, I don't think he'll recognize his own medicine. 
1999/11/17 [Computer/Theory] UID:16906 Activity:nil 
11/16 do people here have opinions on or interest in quantum computing? I am a physicist working on quantum computing, and i'm wondering how many cs people are playing with the idea. at this point, the practical stuff is all still physics, since no one has gotten more than a couple qubits to work, but surely there are more neat algorithms that could be found by cs people in adition to the factoring algorithm( shors algorithm) and the search algorithm(grovers algorithm.) you could argue that this is a waste of time, because we don't know if large scale quantum computation can be achieved, and to that I have no defense(except that we're working on it, and we have a good chance.) i'm just wondering. 
1999/9/24 [Computer/Theory] UID:16585 Activity:moderate 
9.25 Searle will be giving a lecture on the noncomputability of the brain tuesday next week. \_ What time? What room? allenp \_ I'd rather eat bricks. He's an idiot. \_ Just admit that your crushed because he can prove that you cant use that crap you learned in 188 to complete the girlfriend AI project you've been fantasizing about. \_ Possibly. At least I know the difference between "your" and "you're". Anyway, I never took 188 and I got the warez for virtual valerie, so I'm all set in the compuporn dept. Thanks for the thought, though. \_ besides, how many of the sorority girls on campus could pass a Turing test? \_ When I was in school, I had this one roommate in the dorms with the super bimbo sorority gf. I'm pretty sure she couldn't have. They were friends with another sorority chick and dude and I don't think she'd pass either. Nice tits though. 
1999/9/18 [Computer/Theory, Computer/SW/Unix] UID:16544 Activity:nil 77%like:16547 
9/16 How do I pipe to an rsh? I want to do something like sort file  rsh foo l user cat >> bar ? 
1999/9/12 [Computer/Theory] UID:16446 Activity:high 
9.7 For those people who are into AI (or not into it as the case may be): Searle's 9.7.99 lecture for the philosophy of mind class is on: The Chinese Room Argument, and various replies to such. lecture is 9:30  11:00, and 9.7.99 is a thursday @ 2060 VLSB. \_ Meet at 24th and Mission BART. Wear black. \_ I've noticed Searle replies to very insightful questions that could be quite damaging to his theories, by dodging the question and answering just a small portion of the question. If you don't believe me, go yourself, listen very carefully and you will see, he does not quite respond to any question that points to some disagreement with his own theories. \_ You have to remember that the Chinese Room argument is against a very specific claim for AI  not against AI in general. Perhaps that is why you think he is dodging the questions? Then again, Searle's sort of a doofas in general and does that sort of thing... \_ I took his philosophy of mind class a few years ago and suffered through a whole semester of his arrogance and dodging. He's an idiot. \_ Searle's Chinese Room Argument [sic] is an intuition pump. He shows you this system and asks: "What is there that can understand Chinese?" The guy doesn't understand, the paper doesn't understand, the room and the pen don't understand. However, this is just making a claim about individual parts, not the entire system. Certainly no one will claim that individual neurons understand Chinese, yet a system of neurons will. At any rate, what does Searle and the philosophy of mind have to do with AI?  ilyas \_ Searle's class covered AI among other things when I took it. \_ I don't think one can cover AI in a week or two. :)  ilyas \_ Ilyas you're an ass. There isn't a class on campus that can cover an entire subject in one semester you supercilious ass. \_ well, this wasn't meant to be a post for a debate, altho I'm glad some people are talking about things other than ed and unix. It was just an informational post for all our new members who get the motd upon login. Thanks for your opinions tho. \_ JOVE!!!!! \_ _Pico_, buddy. http://www.newgrounds.com/pico \_ EX! \_ /bin/cat > filename!!! 
1999/8/1113 [Computer/Theory] UID:16296 Activity:moderate 
8/11 On the off chance that anyone is familiar with such things I'm having a hard time trying to understand the implementation of the algorithm for division of large numbers presented in Knuth. This is largely due to an ambiguity of the order of "Set; test for this and that; if true, do something else; repeat test if a third thing is true." (Repeat once or many, and which steps?) Section 4.3.1, Algorithm/Step D3, if you can help. Thanks. calbear \_ Dude, you are one boring man. \_ What's the matter, INDUSTRY MAN? Don't have the brain to understand Academia Man? \_ FYI, it is industry and it's not because it's my area. (If it were, would I need to ask?) calbear \_ Walk on over to Don's office and ask him. Tell him I sent you. He'll know what to do. ali. \_ No one's ever there, but when there is, I'm sure your reputation will precede you, Ali. Anyway, I wrote him about the ambiguity. calbear 
1999/8/68 [Computer/Theory] UID:16263 Activity:high 
8/6 So, i'm just wondering, how important is Math 53 (multivariable calculus  i believe it was 50a in the olden days) to a nongraphics software engineer career. \_ Math is useless. Major in the social sciences. Go to law school. Learn some social skills. You'll make more money and meet better looking people. \_ All math is massively valuable to intelligent people who intend to have careers using thier brain in any respect whatsever. \_ In fact I am of the opinion that an entire second major (in math) is worth the time and effort if you have enough units to spare.  ilyas \_ keep in mind there are all kinds of math, so if you are not into "the calculus" or maybe don't believe in real numbers, there is still all kinds of logic, algebra, and formal calculus that can be useful in later life. friendly intros in the philo, comp sci, cog sci, and linguistics dep'ts exist so you don't have to suffer a grad level math class you didn't mean to try. \_ Spelling looks fine to me. Some of the grammar is spotty. If you doubt the usefulness of a math class, you should probably switch to business, or some other major thot does not require brain usage. \_ Why bother getting a CS degree. Just get a job. \_ It was Math 50B. Math 50A was linear algebra and o.d.e . It may or may not be useful depending on what sort of software you develop. Take it anyway. dim \_ Math 50A became Math 54, Math 50B became Math 53 (I was taking them during the rename). I don't do graphics and don't find the material to be particularly relevant to anything I've done since then, but I feel that math classes like that are good for improving general problemsolving skills (which are very relevant to what we all do as computer programmers). They're also an annoying form of torture, because Berkeley doesn't seem to have any profs left who can teach undergrad math. You either get a good TA and learn from them, or you suffer. \_ You don't need it, even if you do graphics muchandr \_ U must n0t B smairt bc eye red 0n D m0teD dat 2 B smairte eye nede 2 n0e l0tz uv maths s0 eye kan werk for 3defx or da d00m guyz n B k00l! 0nlee maths peeple R smairte!!!1111@@@ \_ I dispute this... if you are doing graphics at anything above the undergrad level, multivariable calculus is a necessity, as are a few other branches of math. But I guess that's not important since the dude doesn't wanna to graphics. Lemme important since the dude doesn't wanna do graphics. Lemme ask this: does he plan on doing any kind of specific programming, or does he want to be an unspecialized (unskilled?) coder? That will make the question easier to answer. blojo \_ It can be useful if you ever have to deal with statistics and/or probability beyond the basic level \_ For what it's worth, I remember 50a as being Linear Algebra, not the same thing as multivariable calculus. There is overlap, but. 
1999/3/2930 [Computer/Theory, Computer/SW/OS/Windows] UID:15650 Activity:high 
3/29 Algorithm question: Is an easy way to figure out the number of days between two given dates? Eg. given 98/11/05 and 98/12/09 as input, the output would be 34 days? Is there a C function that can xlate the date into some absolute number and then do the arithmetic and then covert it back? Thanks! \_ Let me guess. You're trying to implement IfModifiedSince for ee 122 right? What I've used since my early days of programming is time_t time(time_t *t); or 'man 2 time' at the unix prompt. It'll return the number of seconds since Jan 1, 1970. jeff \_ Convert both to Julian dates and subtract. \_ http://wwwmacho.mcmaster.ca/JAVA/JD.html guess you'll have to figure out the algorithm from the JavaScript fcns \_ Or just use mktime(3) and subtract. 
1999/3/2324 [Computer/Theory] UID:15634 Activity:very high 77%like:15625 
3/22 Anyone willing to drive down to Stanfurd on Wednesday? Andrew Wiles is giving a lecture there at 7:00pm. Me and a friend of mine would love to tag along :).  ilyas \whos is the target audience of the lecture. where is it at?psb \general audience. Some auditorium (I don't remember, but the brochure is around the CSUA office). Tickets are $8.50, I think.  ilyas. \_ This is a lecture about... what? Getting your cats and dogs to get along better? \_ "Better Breasts and a Tight Tummy in Ten Days  With Your New Personal Trainer, Andeeeeeeeeeee!!" \_ Isn't this the guy who proved Fermat's Last Theorem? PeterM \_ He's the guy that proved the TiniyamaShimura (sp?) conjecture, which Ken Ribet proved implied FLT. \ TaniyamaShimura and it was in the other order chronologically, but both proofs are quite recent. does anyone know how to explain what a modular form is to people with only 50a/50b math? i've explained a structure of the proof [with a lot of black boxing when it came to stuff i didnt understand] but i had a lot of trouble trying to explain the fundamental idea of the mod form. psb \_ meow meow meow meow meow meow meow meow meow meow meow meow meow meow meow meow meow. meow. meow! baxter \_ i have a math degree from berkeley, have published in math, and dont have a clue what you are talking about(because i'm into geometry, and have only had 113 algebra.) I think you just justified the above post about your obscure posts. \_ Infidel bastard! You have defamed the mighty psb! Your children will grow up to fellate dead goats for money and LIKE IT!! \_Is there much money in fellating dead goats? I also have a math degree from Berkeley and have no idea what psb is talking about. \_ why did I tihnk this was gonna read "and am looking for a job" after the Berkeley part? \_ funny. My brother has a math degree From trinity College, Cambridge Univ. in the UK. He designs multi million dollar financial options and can't balance his checkbook to save his life. math hmm... robin \_ Use a computer to balance your checkbook. \_ Use one to fellate dead goats. 
1998/12/34 [Computer/Theory] UID:15060 Activity:high 
12/2 I'd like to take Computer Vision CS280 next year. However, it requires math 53/54. What's the difference between 53 and 54? Thanks. \_ 53: Multivariable calculus  integrating over multiple variables surface integrals, cool looking 3D pictures 54: Linear Algebra but not in the sense that is useful from any kind of computer graphics work. The only thing you need \_ yo: graphics != vision. I took 280 with malik in 1994. grrrreat class!!! nick to know for, say 184/284, is transformations/translation stuff. differential equations are also covered which is not fun at all. DiffEq is used a lot in EE and E&M (especially EE120) but chances are you're not going to have to know what a Wronskian is. I still don't know but those jack ass math profs (referring to Miller of course) think that all this is really important. The basic fact: only about 5% of what you learn in Berkeley math classes will become applicable to you in the future. \_ I'd be curious to know whose orifice you pulled this statistic out of. i use what i learned in berkeley every day. i have been thankful of having learned at least 90% of it at one point or another. you are an idiot if you don't think knowing diff eqs is important. are you satisfied knowing only how to solve diffeqs with constant coeffs? that solve 80% of the problems you will run into as an EE, but that's about it. or are you just satisifed not knowing how to solve diff eqs at all? you think you've been doing fine without them, but the truth is, people in Bangladesh probably think that their lives are just fine too. here's what you need from linear algebra for cs280: bases, transformations, least squares, rank, singularity, eigenvectors and eigenvalues, orthogonal, orthonormal. That's basically all of the Anton book. ali \_ I never use any math beyond some very basic algebra. I use *none* of the math I learned at Berkeley. This pleases me more than you can imagine. \_ Actually, math beyond basic algebra can be pretty useful (especially some calc used in physics, electrical, engineering, and some other science classes). But the difference between the way a high school teaches and a college does is that in high school they teach you how to solve practical problems whereas in berkeley they're so caught up in theoretical proofs that only math majors really care about. \_ I'm not denying it. I'm just glad I'm not in a position to need any of it. I'm not even using HS math. More like Jr. High. Suits me just fine. YMMV. \_ well, good for you. but honestly, 90% is frikin insane. Maybe you had a better prof than I did but I never had to implement the things I learned from math 54 in any physics 7 or EE classes to the extent that you brag about. I admit having to solve simple diff eq's and integrate relatively simple equations but I never had to do boundry equations, prove the so called "Keith Miller's Equivelance Theorem", find eigen vectors, and do most of that math 54 crap. I'd like to know whose orfice you're pulling that BS out of. \_ I dunno, when I took 54 (then called 50B), I actually used what I learned (when I was awake to learn) a week or two lagged in Physics 7B... \_ ali lives on a higher plane. \_ You are all idiots. I use 0% of what I learned in Cal. All I do everyday is to count the money I make and the # of new concubines I get. And I learned how to do that before I was born. College is useless; education is stupid. Ignorance is strength and selfesteem comes from the cavity of an empty head. \_ Take upper division physics, math, chemistry, or astronomy (probably others in engineering, too) and it will make a lot more sense to you. The lower divison math classes are essential to any career in science or engineering (hence, they are required) and really useful in some other fields, too, like economics. dim \_ You can do just fine if you know math just up to Math 53/54. You can do hot sh*t and make much $ with insight and initiative; you can also do hot sh*t that wouldn't be possible without sophisticated math knowledge, and make much $. You may feel fulfilled in either case, depending on who you are. In either case, you will earn the admiration of people who you think are brilliant and people who you think are idiots, and disapproval and disgust from the same range of persons. jctwu \_ Not really. Remember Ted Kazinski, the brilliant math genius. I don't think anyone liked him. \_ especially the rabbits. jctwu \_ "Math is hard! Let's go shopping!" Barbie \_ amen. \_ Was it Barbie or Malibu Stacy? \_ One, last I checked... 
1998/11/2512/1 [Academia/Berkeley/CSUA, Computer/Theory] UID:15027 Activity:nil 62%like:15048 
11/25 CSUA ALUMNI! Come to the CSUA General Meeting on December 2! There will be a general reunion, and foodP and drinkP afterward. Come to where the elite meet for NPcomplete! \_ someone should keep extra good minutes for those of us alumni living too far away to attend  seidl 
1998/9/79 [Computer/Theory] UID:14558 Activity:high 
9/7 Math question: You know those weird symbolic letters like R (real) Z (integers) C (complex)  what the hell is Q \_ Q is the set of rational numbers. Q for Quotient. jon (wannabe math student) \_ jon, you should come to 104 sometimes.  schoen \_ an omnipotent being... \_ Sush! He can hear you... \_ And _can't_ hear your thoughts? (assuming it really existed) \_ As in the weird dude in the red jumpsuit on Star Trek? 
1998/8/13 [Computer/Theory] UID:14451 Activity:nil 
8/11 Has anyone had any experience with any of the upper div math classes (in particular 113 and 125A)? What would one recommend to fullfill the upper div math requirement? \_ Math 110 (Linear Algebra) is pretty easy. Math 113 is only a little more difficult, but a lot less useful for a CS type. dim \_ Agreed... I was a CSMath doublemajor and I use 110 and 113 a lot more than any of the other upper div math courses. Under no circumstances should you take 104. 128A is useful if you expect to do a lot of numeric programming. mogul \_ 128A has to be the worst math class ever. sooo boring. \_I thought 128a was ok last semester with the Icelandic Prof. Adalsteisson, gave us libs and stuff for C/C++. Demmel is teaching it in the fall and I recommend taking it with him. Avoid Rieffel, boring, and stupid fortran. \_ I'm not a math person, but my math friends always said 113 was a piece of cake and 110 was a really hard weeder course. YMMV. \_ 110 is a more theoretical bend on what was taught to you in 54/50A/whatever. If you didn't do well in lower division, you'll choke in 110. Of course, for the psychotics who want to know where integrals come from, there's always 104. It's a worthwhile course in that it gives you a broader picture of How Things Work, though I doubt you'll use much of the stuff in CS. Take it anyway. :)  tmonroe \_ 104 has been relatively easy this summer. Easier than Math 55 because you don't have to worry about "competition". \_Well gosh, I actually majored in math, and found them all to be pretty useful classes. 104 is possibly the hardest class at Cal. I definitely would not take it just to fulfill a one course req. I thought 110 was very easy, and 113 was not that hard if you were willing to stretch your mind in different directions. 128A was a lot of work, and the math content was kind of low. Probably good for the CS tiein. Are those your only choices, or could you take like mathematical logic, or set theory, either of which would be much more applicable to CS? Or even 170, which they didn't offer when I was there. Grrr. cody \_ Take 104 if you really want to learn math. Otherwise, 110 or 113, if you just want to fulfill a requirement. \_ Math 121a and Math 121b were two of the best classes I took at Cal. I had professor Podles, but some of my friends who took these classes with professor Neu also enjoyed them very much. I also took Math 104 which I think is pretty useless. Even though Math 104 is interesting and you learn some fancy words like "compact", the only benefit I recieved was learning how to write a proof (e.g. I leaned what an upside down A and a backword E mean). Feel free to email me if you would like to read more of my rant on why Math 121a and Math 121b are FAR superior to math 104, 110, 113, 185 and most others with the possible exception of Math 128a/b or the applied math course (I don't remember the course number for the applied math course). By the way I got an EE/CS degree with a minor in physics, but I am working as a software engineer now. emin \_ Which 12 step program did you attend? \_ You used to be able to satisfy the L&S upper div math requirement with stats 134, which was a basic intro to probability and statistical methods. not a bad thing if you're not into most mathematicians' math and you feel a little too focused towards discrete math and logics in CS karlcz 
1998/8/13 [Computer/Theory] UID:14447 Activity:nil 
8/12 Apparently you can now patent math formuls. http://www.npr.org/programs/atc/archives/1998/980811.atc.html  seidl 
1998/7/26 [Computer/Theory] UID:14396 Activity:nil 
7/26 CK Miller, the worst math teacher I've ever had in my life. If you're taking math 53 with him this fall be scared. BE VERY SCARED. 
1998/6/1113 [Computer/Theory] UID:14205 Activity:kinda low 
6/11 Why the hell is the new RSA command line all screwed up???s \_ What are you talking about? \_ RSA has a command line? I thought it was a cryptography method. \_ The secret backdoor that the NSA/CIA/FBI/BATF put into the RSA algorithm means that they can extract anyone's private key from any encrypted data by running a simple (but highly secret) Perl script over that data from the command line. 
1998/4/910 [Computer/Theory] UID:13927 Activity:very high 
4/9 What algorithms are patentable and what algorithms are not? (ie. RSA is, quick sort isn't wtf?) \_ uh, are you familliar with the concept of prior use? \_ none should be League for Programming Freedom \_ It all depends on whether you can get a court to agree with you that an algorithm is (or isn't) patentable. 
1998/4/6 [Reference/BayArea, Health, Computer/Theory] UID:13907 Activity:nil 
4/6 Re: PBS starts new series > Teletubbies \_http://www.sfgate.com/cgibin/article.cgi?file=/chronicle/archive/1998/04/06/DD91651.DTL 
1998/4/4 [Computer/HW/Memory, Computer/Theory] UID:13899 Activity:nil 
4/3 How do you solve: 1) summation of 1/n where n = (1 to infinity) 2) summation of 1/n^2 where n = (1 to infinity) THIS IS A SERIOUS QUESTION, THANKS!!! \_ If you sum from 1 to infinity instead, the first one, at least, still diverges to infinity. \_THANKS! How do you solve by induction w/base case? \_ Is your homework due at 5 PM? \_ Already due. Actually I'm at work and I don't have my calculus book. \_ I'm trying to solve the interview question: How do you write a program that can randomly and evenly pick ANY line in a huge text file, in one parse, where you have very little memory? 
1998/3/2527 [Recreation/Computer/Games, Computer/Theory, Recreation/Sports] UID:13862 Activity:high 
3/25 Are there any practical applications of Conway's Game of Life? \_ xlock \_ [ignorant babbling about the board game deleted] \_ A PC moria/angband/rogue/etc clone called ADOM uses it to determine growth for various plant life in the game. \_ Just saying "no" would have been shorter. \_ I thought there was some use for that and other cellular automata in generating pseudorandom numbers. \_ there's an encryption scheme based on a 1D cellular automaton called CA with a standing challenge to crack it \_ cellular automata are still an inspiration for some artificial life work. And they're equivalent to Universal Turing Machines, so you can do math on Life if you want to go to the trouble. 
1998/3/1920 [Computer/Theory] UID:13841 Activity:high 50%like:13522 
3/20 Would someone please recommend a good book on Black Scholes Option Pricing Model? \who is asking? \_ John C. Hull's book _Options, Futures, and Other Derivatives_ does a pretty good job  android \_ duffie's book also does the job, but hull's is better. fab \_ Beware that some of the Black Scholes formulas have been shown to be inaccurate. Look for Robert C. Merton's \this is a deep philososphical matter. psb books as well. I believe he actually did the math for Black/Scholes, despite Merton's objections about the errors in some of Black/Scholes eqns. I think Merton's book is called "continuous time finance", 92. \David Cox [the finance one of from stanford, not the math guy from i think columbia] has some fine mathematical finance books. psb \i know i am going to get it for this, but i am having dinner on sat with a friend of the family who was invited to merrilllynch by scholes. if you want me to ask him a specific question, send me some mail asap. psb 
1998/2/1011 [Computer/Theory] UID:13653 Activity:high 
2/10 size(NT5) = 2 * size(NT4). Oh boy. \_ Twice the bloat, half the power, quadruple the bugs. Although I'm pretty sure that size(NT?) will overflow anyway so this isn't a valid formula on most computational devices. Might be computable if you had a *lot* of paper. \_ Isn't quadruple rather optimistic? \_ I was figuring it'd be quadruple after service pack 4 or 5. 
1998/2/811 [Computer/Theory] UID:13640 Activity:kinda low 
2/8 What does "complied with RSAREF" (or does not use RSAREF) mean? \_ See "http://www.rsa.com/rsalabs/newfaq/q174.html" It's the "RSA Reference implementation". There are involved political and engineering questions surrounding its use or nonuse, mostly because it's kind of slow  but the algorithm is protected by a patent in the US, so you might get sued if you try to write your own faster or more specific RSA code for certain applications without a license from RSADSI. \_ If it uses RSA in the US, you need to use code licensed by RSA. For noncommercial software, that means RSAREF or nothing. Anything else is illegal. (If you think software patents like this are stupid, go to http://www.lpf.org and join them.) \_ Hey, this url doesn't work. No DNS entry, friend. Got a mirror or something? \_ Weird. Guess RMS didn't pay the bill. Anyways, try ftp://prep.ai.mit.edu/pub/lpf/ for now \_ It's all kinda old. Is this org still active? 
1998/1/21 [Computer/Theory] UID:13537 Activity:high 
1/20 math 110, dubins or frenkel? comments on either appreciated. jon \_ dubins is starting to go senile \_ No math class is worth taking unless it's being taught by the _man_ himself, H.H. Wu! \_ They still let Wu teach? How about the French guy, Pew? (sp) 
1997/4/1 [Computer/Theory] UID:32118 Activity:nil 
3/31 Someone once told me the story of some graduate math student who got to class so late that the class was over and there was some unsolved math theorem on the board. he thought it was home work and went home and solved it, and became famous. does anyone know if this is true? kane \_ it sounds apocrophal \_ It is a true story. The student was George Dantzig  android. \_ And what was the problem? \I dont know the problem, but the answer was the Simplex Method. psb \_ actually, i heard a similar story about John Milnor which goes like this: Milnor was in an undergraduate class at Princeton and came in late the prof had put an unsolved problem on the board and Milnor thought it was homework the problem was: is there a minimum turning angle for a nontrivial knot Milnor went home and proved that a nontrivial knot most turn though at least 4*pi (two revolutions) and so he published his first result at 18 years old kinda makes the rest of us mortals sick jkin \_ And I heard a similar story about someone different. I think the blackboard /homework part is apocrophal (or at best, exaggerated) even tho the underlying problems and people are real. \_ http://www.urbanlegends.com 
1994/4/28 [Computer/Theory, Computer/SW/Unix] UID:31579 Activity:nil 
4/27 mailagent sort of installed. have fun reading the man page: psb soda{400}[/usr/local/man/man1]% ls s  sort nr 176 mailagent.1 112 gcc.1 95 trn.1 
1993/12/12 [Computer/Theory] UID:31446 Activity:nil 
11/13 People wishing to be involved in a sort of combative automatatype contest not dissimilar to Core War which will be taking place in Texas on November 30 should contact me PDQ. It looks to be a really classy gathering with guest speakers and stuff. Team of 3 undergrads. blojo 
1993/12/10 [Reference/BayArea, Computer/Theory, Computer/SW/Unix] UID:31442 Activity:nil 
12/7 There is some sort of showing of a work in progress of a movie of the book 'Neuromancer' at Mills at 9pm on Friday. Anyone interested in going with me? danh \_ Currently, we're thinking of pub trans... anyone drive? marco \_ sparky's \_ anyone willing to risk the Payammobile can go with me \_ 1 space left \_ It is showing at 9pm at Mills College Concert Hall, 5000 MacArthur Blvd., Oakland, CA. Supposedly AC Transit bus lines 57 and 58 N stop directly in front. mail snol@mills.edu for even more detailed info. danh \_ Mebbe we should meet someplace in Berkeley so the AC Transit clueless (like myself) don't get hopelessly lost in Oakland. sameer \_ I can pick up and take 12, or 3 really uncomfortably but, it's free and not that long a ride. thorn \_ I'd be willing to drive a load out, too. ERic \_ Anyone have enough room to give me and my email babe a ride?  danh \_ How fat is she? \this a virtual babe? \_ I might be able to make it. If so, I can drive people. phr \_ if i can't find a spot in a car, i might offer a spot on the back of my Suzuki. hh 
9/18 