As the change from 1.4 to 1.5 db will be a big step anyway, two additions come to mind:
- Language links
Should we finally put these into a real table? At least, additionally to keep them in the text? With an interwiki link table up and current, we could then switch to "real" interwiki management at a later stage.
Have there been any considerations to add support for the following ideas: http://meta.wikimedia.org/wiki/Reviewed_article_version
Schewek
On Tue, 29 Mar 2005 03:30:25 +0800, b schewek schewek@linuxmail.org wrote:
As the change from 1.4 to 1.5 db will be a big step anyway, two additions come to mind:
- Language links
Should we finally put these into a real table? At least, additionally to keep them in the text? With an interwiki link table up and current, we could then switch to "real" interwiki management at a later stage.
Have there been any considerations to add support for the following ideas: http://meta.wikimedia.org/wiki/Reviewed_article_version
People have not yet agreed on [[meta:Article_validation]] - even some basic issues, such as whether to show to anons the reviewed or latest version. I can't imagine support for this will be in 1.5.
Tomer Chachamu (the.r3m0t@gmail.com) [050329 06:30]:
On Tue, 29 Mar 2005 03:30:25 +0800, b schewek schewek@linuxmail.org wrote:
Have there been any considerations to add support for the following ideas: http://meta.wikimedia.org/wiki/Reviewed_article_version
People have not yet agreed on [[meta:Article_validation]] - even some basic issues, such as whether to show to anons the reviewed or latest version. I can't imagine support for this will be in 1.5.
That's a really bad thing, considering a pile of plans are hung waiting for just that feature or something like it.
The current suggestion is to try running ratings and see if they end up with sensible results. So rather than using the results, we just ... gather the data then the data is sanity-checked. This would mean making sure the feature was *stable* (production-ready) rather than *ideally feature-complete*, because what's done with any ratings would almost certainly change.
Would a test version of the feature (that gathers data but does not do things with it) be more of a prospect for 1.5?
- d.
David Gerard wrote:
Tomer Chachamu (the.r3m0t@gmail.com) [050329 06:30]:
On Tue, 29 Mar 2005 03:30:25 +0800, b schewek schewek@linuxmail.org wrote:
Have there been any considerations to add support for the following ideas: http://meta.wikimedia.org/wiki/Reviewed_article_version
People have not yet agreed on [[meta:Article_validation]] - even some basic issues, such as whether to show to anons the reviewed or latest version. I can't imagine support for this will be in 1.5.
That's a really bad thing, considering a pile of plans are hung waiting for just that feature or something like it.
Yes, I think we need to have an understanding of what needs to be done in this 1.5 version. We do *not* need to decide such issues as whether to show anons the reviewed or latest version. All we need for starters is... starters.
What I envision for 1.5 is the simplest possible data gathering tool.
People rate articles, and we record everything about that rating -- who did the rating, what was the rating, what version was rated, etc. Then we do nothing at all with the data except study it. We can anonymize it and share it, and people can run studies of various kinds on how to combine the ratings effectively. We can look at the ratings and see if they are sane, or where they are sane... do anons do a good job of rating? do experienced editors tend to give the same ratings, etc.
The beauty of just gathering data and studying it for awhile, with *no actual implications for the site*, is that we don't have to "a priori" figure out how to prevent the ratings system from becoming a slashdot-style karma-whoring game. We just gather the data, and think really really hard about it.
--Jimbo
On Mon, 2005-03-28 at 14:34 -0800, Jimmy (Jimbo) Wales wrote:
What I envision for 1.5 is the simplest possible data gathering tool.
Agreed...it's the data that's important, and the presentation can wait. But it would sure be nice to have some sort of general facility for metadata in place, and a better facility for user identification and validation (for example, what would happen if we collected all this data about who likes what articles, and then someone hijacked the user accounts of the reviewers and changed their data?) Some sort of basic signature system would solve that pretty easily.
It should also be pointed out that there's no reason metadata has to be in the Wiki database itself, as long as there's a stable and standard way to /point/ to an article revision (another thing in dire need of standardization). The wiki software should probably try to facilitate entering the data, though.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Jimmy (Jimbo) Wales schrieb:
Yes, I think we need to have an understanding of what needs to be done in this 1.5 version. We do *not* need to decide such issues as whether to show anons the reviewed or latest version. All we need for starters is... starters.
What I envision for 1.5 is the simplest possible data gathering tool.
People rate articles, and we record everything about that rating -- who did the rating, what was the rating, what version was rated, etc. Then we do nothing at all with the data except study it. We can anonymize it and share it, and people can run studies of various kinds on how to combine the ratings effectively. We can look at the ratings and see if they are sane, or where they are sane... do anons do a good job of rating? do experienced editors tend to give the same ratings, etc.
The beauty of just gathering data and studying it for awhile, with *no actual implications for the site*, is that we don't have to "a priori" figure out how to prevent the ratings system from becoming a slashdot-style karma-whoring game. We just gather the data, and think really really hard about it.
Done. :-)
(In case you didn't notice, I re-implemented the validation feature in 1.5, and Brion wants to run a demo on his own server soon.)
Magnus
On Tue, 2005-03-29 at 03:30 +0800, b schewek wrote:
Have there been any considerations to add support for the following ideas: http://meta.wikimedia.org/wiki/Reviewed_article_version
Something along those lines is pretty high on my priority list, but awaits better handling of metadata in general.
wikitech-l@lists.wikimedia.org