Path: news.mathworks.com!not-for-mail
From: "John D'Errico" <woodchips@rochester.rr.com>
Newsgroups: comp.soft-sys.matlab
Subject: Re: Doubt on usefulness of the FEX
Date: Sat, 8 Nov 2008 02:16:02 +0000 (UTC)
Organization: John D'Errico (1-3LEW5R)
Lines: 60
Message-ID: <gf2sp2$141$1@fred.mathworks.com>
References: <gev97h$pim$1@fred.mathworks.com> <gevjr0$mfl$1@fred.mathworks.com> <gf16vq$nsn$1@fred.mathworks.com> <gf1cjp$jfh$1@fred.mathworks.com> <gf2bng$9dq$1@fred.mathworks.com>
Reply-To: "John D'Errico" <woodchips@rochester.rr.com>
NNTP-Posting-Host: webapp-05-blr.mathworks.com
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: 8bit
X-Trace: fred.mathworks.com 1226110562 1153 172.30.248.35 (8 Nov 2008 02:16:02 GMT)
X-Complaints-To: news@mathworks.com
NNTP-Posting-Date: Sat, 8 Nov 2008 02:16:02 +0000 (UTC)
X-Newsreader: MATLAB Central Newsreader 869215
Xref: news.mathworks.com comp.soft-sys.matlab:499624


"James Tursa" <aclassyguywithaknotac@hotmail.com> wrote in message <gf2bng$9dq$1@fred.mathworks.com>...
> "John D'Errico" <woodchips@rochester.rr.com> wrote in message <gf1cjp$jfh$1@fred.mathworks.com>...
> > 
> > Yes. I've been disappointed in the FEX mods.
> > I think a very useful thing would have been
> > to filter out stuff. If you never want to see
> > anything by Marco, then this should be trivial
> > to do. Somehow I understood something like
> > this was to have been part of the new FEX.
> > For example:
> > 
> > I'd like to be able to assign some weight to
> > any individual, say in the range [0,1000].
> > Anything they say or do on the FEX is applied
> > this weight. Define all default weights for
> > every person as 100. So for Marco, I'd reduce
> > his weight to 0, so that nothing he does will
> > ever even show up.
> > 
> > If I'm looking at a file, and Marco has made
> > some ridiculous comment on it, that comment
> > is itself filtered by the same weighting system.
> > 
> > On the other hand, a person I respect would
> > be given a higher weight, so that their works,
> > even their comments, would naturally float to
> > the top. For some reason, I had been led to
> > believe that this (or something like it) was to
> > have been part of the new FEX mods.
> > 
> > John
> .
> In addition to personal filtering or ranking, that is only visible to the individual, I would also like to (selfishly) be able to see these rankings from other selected users. For example, if I trust yours & Jos's & other specific individual's rankings, and if you have already done some work evaluating certain submissions and ranked them highly, I would like to know that and use that information to determine which submissions are worthwhile for me to download. So some method of combining rankings only from other selected individuals would be useful to me. Even a simple list of Top 10 favorite downloads, or Top 10 favorite posters, from the experts would be useful.
> .
> Also, the current weighting system is often not very much help. I would rather see a scoring system of some sort that takes into account the number of high rankings, not just an average. If one person rates a M___ (I can't bring myself to spell it) submission a 5, then it shows up as a 5. And the more crap he submits, the more downloads he gets and that just feeds the troll more.
> .
> It is almost like we need a subsection of the FEX, and the only way you get into it is by multiple peer review from approved reviewers, or by amassing a minimum number of high ranks from  the general population.
> .
> James Tursa
> .

We tried to make some of these ideas work.

For example, we tried selecting an upper
subset, what we were going to call "Select".
It died off due to disagreements in terms of
exactly what qualified as a Select code before
we got it going though. This is filtering in a
sense, by using a team of "trusted" reviewers
to identify a set of really good codes.

Another idea I tried once was to work through
every single submission in one of the old FEX
categories. I downloaded every file, read
through it, then tried to rate it on a list of
standard merits. It was a huge amount of
work though, to do this for a few hundred
files. 

John