On Jan 22, 2008 5:14 PM, Anthony
<wikimail(a)inbox.org> wrote:
On Jan 21, 2008 10:33 PM, Ian Woollard
<ian.woollard(a)gmail.com> wrote:
On 21/01/2008, David Goodman
<dgoodmanny(a)gmail.com> wrote:
this seems a little circular.
It does *seem* so, but it's not circular, since nobody is made notable by
noting themselves.
how do we tell who is notable in the first
place?
Because other people note them in turn. ;-)
and how to we get out of this trap?
If you think about it, this is the same kind of problem faced by search
engines. When you do a search for web pages they give you what we can call,
for the sake of this argument, the 'most notable' web pages that contains
the words you're looking for, where notability is related by how many web
pages link to a page, and how many link to the pages that link to them, and
so on.
Which is the same thing, And it's a solved problem.
So in principle the same formal algorithms (e.g. PageRank) can be applied to
the wikipedia concept of notability (but of course notability in this case,
not over webpages, instead over all the books, films, magazines, people's
comments etc. etc.) And we would get an unambiguous number that corresponds
to notability.
Seems to me that would correspond more to popularity than to
"notability". These two concepts are different, right?
And what's the cutoff which qualifies as "notable enough"? Google's
PageRank works because there's no cutoff. If I type in "Wikipedia", I
get the page with the highest PageRank for "Wikipedia". But if I type
in "Capriccio" I get the page with the highest PageRank for
"Capriccio". I don't get a message saying "Sorry, no links for
'capriccio' are notable enough".
Of course in the real world, we aren't
running the algorithm, and we expect
that editors to more or less know who and what are notable and who aren't,
and it may look very different at first. But I think if you look at what the
people are doing, it amounts to essentially the same idea as what google do
with webpages; but run in peoples heads in a distributed way, they keep
track of the most notables for the subjects they are interested in in much
the same way.
Right?
I think some people are treating notability that way, and I think this
comparison to Google is a good example of why it's such a bad idea.
_______________________________________________
WikiEN-l mailing list
WikiEN-l(a)lists.wikimedia.org
To unsubscribe from this mailing list, visit:
http://lists.wikimedia.org/mailman/listinfo/wikien-l
Google is meant to be indiscriminate and to index everything (or at
least everything which a robots.txt or the like doesn't explicitly ask
it not to). Wikipedia is explicitly not intended to be so.
So, what do we ask? "How much independent reliable source material is
available on this subject?" If the answer is "a large amount", we
write a full article. If the answer is "a little bit", we might have a
suitable list or parent article to make a mention. If the answer is
"none", we write nothing. All of these are in keeping with our core
policies. "No original research" indicates that we don't second guess
reliable sources, just follow their lead. If their lead is "this isn't
important enough to write about", we don't say "Well they're
wrong",
we follow their lead-and don't write. Undue weight indicates much the
same thing-sources indicate how much weight we give something. If they
decide "very little", we write very little. If they decide "none",
we
write nothing. We don't decide, reliable and independent sources do.
Period. That's our metric. We don't need any other. Nice, simple, and
no need for editors to decide at all.
As to determining reliability, this is a solved question. Nature and
Science are reliable sources. The Weekly World News is not. The New
York Times is a reliable source. A tabloid sensationalist rag is not.
In some cases there might be an edge case, but in most cases we can
use simple metrics. Is the source widely regarded as reliable? Is it
written, peer-reviewed, editorially controlled, and/or fact-checked by
professionals? Is it cited by other sources known to be reliable?
Again, we let others decide, we don't need to do so ourselves, and we
-shouldn't- be doing so ourselves.
The above is, in my opinion very over-simplified. It is also subjective.
Yes, we can recognise really good reliable sources and really bad
reliable sources, but most of the time we are in the middle.
The problem really is thinking that sources determine whether we write
about something, rather than needing sources to write. A current Afd is
interesting:
The keeps are saying there are sources. The deletes are saying, yes, but
it is not encyclopedic. I !voted keep as the notability guideline agrees
with you. If there are good sources we have an article. However, I still
do not think this article is appropriate. We need to improve the
inclusion criteria so we include what is worthy of being in an
encyclopia and reject stuff that is not worthy. Just going for sources
leads to lots of crap on wikipedia and perhaps to rejecting stuff that
does not have conventional sources.
We need to think hard about this and not just accept over-simplified
solutions.
Brian.
--
Freedom is the right to say that 2+2=4. From this all else follows.
_______________________________________________
WikiEN-l mailing list
WikiEN-l(a)lists.wikimedia.org
To unsubscribe from this mailing list, visit:
http://lists.wikimedia.org/mailman/listinfo/wikien-l
--
Brian Salter-Duke b_duke(a)bigpond.net.au
[[User:Bduke]] mainly on en:Wikipedia.
Also on fr: Wikipedia, Meta-Wiki and Wikiversity