[Foundation-l] "RevisionRank": automatically finding out high-quality revisions of an article

[Yao Ziyuan]

On Tue, Dec 20, 2011 at 9:07 PM, WereSpielChequers <
werespielchequers at gmail.com> wrote:

> >
> > ------------------------------
> > On Tue, Dec 20, 2011 at 6:38 AM, Yao Ziyuan <yaoziyuan at gmail.com> wrote:
> >
> > > Hi Wikipedians,
> > >
> > > I seem to have found a way to automatically judge which revision of a
> > > Wikipedia article has the best quality.
> > >
> > > It's very simple: look at that article's edit history and find out,
> > within
> > > a specified time range (e.g. the past 6 months), which revision
> remained
> > > unchallenged for the longest time until the next revision occurred.
> > >
> > > Of course there can be additional factors to refine this, such as also
> > > considering each revision's author's reputation (Wikipedia has a
> > reputation
> > > system for Wikipedians), but I still feel the above idea is the
> simplest
> > > and most elegant, just like the original PageRank idea is for Google.
> > >
> > > Best Regards,
> > > Ziyuan Yao
> > >
> > >
> >
> >
> > ------------------------------
> >
> > Message: 11
> > Date: Tue, 20 Dec 2011 01:16:15 +0000
> > From: Tom Morris <tom at tommorris.org>
> > Subject: Re: [Foundation-l] "RevisionRank": automatically finding out
> >        high-quality revisions of an article
> > To: Wikimedia Foundation Mailing List
> >        <foundation-l at lists.wikimedia.org>
> > Message-ID:
> >        <CAAQB2S9TEQhuiaD4Gb0ZjV-tVSLRgmmjHjbHJwAww=
> 8UiFtf3A at mail.gmail.com
> > >
> > Content-Type: text/plain; charset=ISO-8859-1
> >
> > On Mon, Dec 19, 2011 at 22:38, Yao Ziyuan <yaoziyuan at gmail.com> wrote:
> > > I seem to have found a way to automatically judge which revision of a
> > > Wikipedia article has the best quality.
> > >
> > > It's very simple: look at that article's edit history and find out,
> > within
> > > a specified time range (e.g. the past 6 months), which revision
> remained
> > > unchallenged for the longest time until the next revision occurred.
> > >
> > > Of course there can be additional factors to refine this, such as also
> > > considering each revision's author's reputation (Wikipedia has a
> > reputation
> > > system for Wikipedians), but I still feel the above idea is the
> simplest
> > > and most elegant, just like the original PageRank idea is for Google.
> > >
> >
> > Okay, how about this.
> >
> > I find a page today that has had only one edit in the past year. That
> > edit was an IP editor changing the page to insert the image of a man
> > sticking his genitalia into a bowl of warm pasta (I haven't checked
> > Wikimedia Commons but would not be surprised...).
> >
> > Nobody notices the change until I come along and undo it. I then see
> > that it is a topic that interests both myself and a friend of mine,
> > and we collaborate on improving the article together: he writes the
> > prose and I dig out obscure references from academic databases.
> > Between us, we edit the page four or five times a day, every day for a
> > week improving the article until it reaches GA status. Having
> > nominated it for GA, a WikiProject picks up on the importance of the
> > topic and a whole swarm of editors interested in the topic swoop in
> > and keep editing it collaboratively for months on end.
> >
> > Under your metric, in this scenario, the edits of a sysop and an
> > experienced user, or later the WikiProject editors, would not be
> > chosen as the high-quality stable version.
> >
> > As for author reputation, check out the WikiTrust extension for
> > Firefox - see http://www.wikitrust.net/
> >
> > --
> > Tom Morris
> > <http://tommorris.org/>
> >
> >
> >
> >
> Hi Ziyuan Yao, that is an interesting idea, but not necessarily something
> that one should do automatically.
>
> I recently found an article that since 2006 had been telling the world
> where the Holy Grail had been from the closure of Glastonbury monastery
> until the start of the twentieth century. It will take some years of
> non-editing for the new version of that article to become the stable one.
>
> Also some of the articles that our readers are most interested in would
> look a tad dated. Sarah Palin's article may no longer be at the 25 edits
> per minute stage that it peaked at, but how many years will it be before it
> becomes as stable as it was a week before she became John McCain's running
> mate?
>

First, read my last message mentioning Debian.

It's possible that we give different types of articles different periods of
time for trial. We can manually specify that articles under a certain
category must wait 2 months to be considered mature while articles under
another category must wait 1 year. Debian also allows software testers to
set different waiting time for each package (called "urgency"); urgent
packages can graduate faster than normal packages.

Or we can automatically determine the waiting time for each article based
on how hotly it is edited. Hotly edited articles like Sarah Palin can
automatically have a shorter waiting time to become mature. It's all
relative to an article's editing frequency and viewing frequency.


> Of course the edit history is out there so the earlier versions are
> available under the same license as the current version. So any
> enterprising mirror could adopt a system like this if they thought it would
> look at least as good as the current Wikipedia. As far as I know no-one has
> yet, and I suspect if they did they'd have legal problems re libellous
> statements about living people. Wikipedia at least has the moral and I hope
> legal defence that when we learn of an error we fix it. This sort of system
> would be automatically displaying an earlier version despite knowing that
> in many cases it would be displaying false and damaging information.
>

I'm less concerned with legal problems. I think maybe they can just put a
disclaimer before every article, saying "This article was last updated
MM/DD/YYYY and we don't guarantee its accuracy."


>
> WSC
>
> WSC
> _______________________________________________
> foundation-l mailing list
> foundation-l at lists.wikimedia.org
> Unsubscribe: https://lists.wikimedia.org/mailman/listinfo/foundation-l
>