Hey guys,
I've read a lot on the whole wikidata thing in the past view month. The whole project will IMO be a enourmous advantage not only to mediawiki and wikipedia but also to the whole idea of free information.
I think, however, that most of you concur with me in this point.
There are loads of amazing ideas in the minds of all us and in meta.wikimedia.org, so this is not the point ;)
What the project lacks of at the moment is a better coordination and organisation.
I think that we should link the idea of wikidata with some other improvements that mediawiki needs, especially better semantic web abilities, iow. a xml in- and output implentation, the possibilty to tag articles with a standart (as for Jimbo's "1.0"-proposal) and maybe also a improved discussion system (liquidthreads).
So I propose to take all this together and call it MediaWiki 2.0 resp. phase4. I know, the idea of MediaWiki 2.0 existed all the time as a "dream for the future" with no atcual plans for the moment.
But why not start working more target-aimed? When to start, when not now?
Even if you think that this is too much for one task, IMO we should do this just for wikidata as well.
So I propose to do the following:
- create a newsgroup "gmane.org.wikimedia.mediawiki.wikidata" or "gmane.org.wikimedia.mediawiki.2-0" to discuss the stuff there instead of on [[talk:wikidata]] - and, which is very important, decide ASAP how to implent wikidata into the database, there are quite a few models flowing around in meta.wikimedia.org (e.g. after a last regarding of pros and cons of the different models in the newly created newsgroup ;) ) - decide ASAP in which areas of the current mediawiki code major changes are needed (e.g. if we want to rewrite the parser to support different output formats natively or if we just include a xml/rdf output abilty into the existing code
when all this is and further discussion is done we should
- create a new cvs branch 'mediawiki20' (or use the wikidata tag that does already exist) or a new module 'phase4' - create a roadmap which shows the *concrete* steps towards the realisation
I think we have enough guys out there who're more than willingly to help coding, but the problem is that no one really knows where to start ..
Concerning myself, I think I will not be able to write too much of code myself, because my little freetime is already rather occupied, but I'd love to contribute to the project as much as possible.
best regards, Frando
On 11/7/05, Frando frando@xcite-online.de wrote:
I've read a lot on the whole wikidata thing in the past view month. The whole project will IMO be a enourmous advantage not only to mediawiki and wikipedia but also to the whole idea of free information.
I think, however, that most of you concur with me in this point.
There are loads of amazing ideas in the minds of all us and in meta.wikimedia.org, so this is not the point ;)
[snip]
This is a complex subject. If we try to discuss everything out we'll just be in discussion forever. It's pretty much come to a point in time where we should say "Show me the code".
If it's going to move forward it will either be via incremental improvement, or via a crazed rewrite (if you do this, please write the parser in a c-callable language and make it versatile enough to use it outside of mediawiki!).
Either approach is valid it's just a question of who is motivated to do what work.
Frando wrote:
What the project lacks of at the moment is a better coordination and organisation.
Erik Moeller is working on it; he *is* the project, and I am confident that he can coordinate himself :-)
I think that we should link the idea of wikidata with some other improvements that mediawiki needs, especially better semantic web abilities, iow. a xml in- and output implentation,
My initial but mostly working XML export implementation:
http://magnusmanske.de/wikipedia/wiki2xml.php
Paste wikitext in, get XML out. I've been meaning to work on a further conversion to OpenOffice XML, but didn't really get around to it.
the possibilty to tag articles with a standart (as for Jimbo's "1.0"-proposal)
It's waiting in the wings, for Brion to copy SpecialValidate.php back from CVS HEAD and turn it on.
and maybe also a improved discussion system (liquidthreads).
Didn't write that yet. Surprise! ;-)
So I propose to take all this together and call it MediaWiki 2.0 resp. phase4.
Why? They're not dependent on each other in any way. WikiData is developing, validation is ready for testing; XML export(/import) is in early stages, and LiquidThreads is in consensus limbo. None of them needs any of the others to become reality.
We can say "together, we'll call them MediaWiki 2.0", but that's a marketing thing, not a technical necessity IMHO.
I think we have enough guys out there who're more than willingly to help coding, but the problem is that no one really knows where to start ..
* WikiData should stay with Erik for the time being; in such an early phase, a single developer can be more effective than a dozen guys who have to be told every detail first. We'll grab it once it has reached a beta or whatever * Validation waits for field testing, though I'd be happy if you'd have a look at the code * XML import/export would be something where you can jump right in. Speed up/bugfix my parser, add an ODT converter, write an importer (easy enough), whatever * LiquidThreads seems to be stuck between better order and security vs. "everyone can edit this comment".
Concerning myself, I think I will not be able to write too much of code myself, because my little freetime is already rather occupied, but I'd love to contribute to the project as much as possible.
Welcome to the club of people-who-don't-have-time-but-still-make-some :-)
Magnus
wikitech-l@lists.wikimedia.org