10/20 Hi, I've created a toy web site that will hopefully be a bit of
insightful for people who want to know the "slant-ness" of different
news source: http://www.slantcheck.org
I already bought the domain names, I'm now looking for a place
to host it. If you would like to help please email me. -kchang
\_ http://www.free-webhosts.com/webhosting-01.php
\_ Kevin, does it occur to you that averaging faulty sensor
readings doesn't produce meaningful results? Maybe if we
had some sort of pagerank for people this could work. -- ilyas
\_ the same is true for web votes on http://cnn.com, http://cbsnews.com, etc.
Also read his disclaimer. It's not meant to be scientific at all
\_ I know. I am saying why add to the garbage? -- ilyas
\_ ilyas-- what is trash to you may be useful to others.
To say categorically that something has no value,
says a lot about you. Secondly, most systems require
some level of trust and certainly all systems are
subject to abuse. Just look at the electoral college,
Gerrymandering, e-vote machines crashing, etc.
No system is abuse free -- some systems are much
more abuse prone than the others (case in point informal
internet vote). It's good to have a starting point
somewhere, and in time, refine the system to a point
that it is much less abuse prone and that it is
generating acceptible results.
\_ It does say a lot about me. It says that I think
systems where a vote is trivial to fake, where
a single person can trivially cast arbitrary
numbers of votes, where the opinions of all
people are weighed equally, etc. etc. etc.
will produce garbage. No one will
rely on such a system for anything other
than generating empty motd conversations.
Having said that, I welcome differing opinions
of 'others,' because I am curious how http://cnn.com
polls can possibly be of any use to anyone.
I want to be proven wrong here. If you honestly
want to make progress in this area, you can
look at social networks/pagerank research,
or computer security. -- ilyas
\_ Aw, I thought it was going to run news articles through some sort
of analysis program to compute the results. Instead I find it's
just an unfiltered click poll.
\_ that itself is a PhD thesis right there. Context sensitive
weight analysis.
\_ Yeah, well I could hope for some arbitrary heuristics at
least. A poll isn't right... the name evokes http://factcheck.org
which at least provides human analysis. A <DEAD>slantcheck.org<DEAD>
run by some dedicated individuals who analyze submitted
instances of "slant" could actually be an interesting
service that could get national attention.
\_ Is this thing just a cry for attention?
\_ I dunno. But a http://factcheck.org comparison is natural...
hey I would enjoy doing that analysis as part of
some funded group. Those http://factcheck.org people get
paid to sit around and analyze the same shit you
guys all do on the motd every day.
\_ thanks for the response guys. The bottom line is that there are
a lot of improvements and changes that need to be made in
order to make the results fair and meaningful. I'd love to
implement some of the features that were suggested, but most
of them require a lot of time and/or money. Please keep up these
great suggestions, but even more importantly, send me money
via PayPay. Once I generate enough interests and funding,
I'll be able to hire someone to implement these
features. Thanks. -kchang
\_ How are we supposed to know you aren't going to spend it all
on h07 42n ch1x, or hire one to "implement" your features?
\_ he's gonna hire hot UCLA chicks to implement the features :) |