On Mar 5, 2007, at 1:10 PM, Rob Church wrote:
On 05/03/07, Simetrical Simetrical+wikilist@gmail.com wrote:
On 3/5/07, Jim Wilson wilson.jim.r@gmail.com wrote:
I doubt most Wikimedians even know that the inclusion of new extensions would be considered. I certainly didn't - until you just said something.
Is there a sounding-board somewhere that people can make these suggestions and the community can discuss it?
As far as I've seen, it's mostly harassing Tim and/or Brion until they give you an answer.
You oversimplify too greatly.
A huge proportion of the extensions out there are written by, if I may be frank, amateurs.
Indeed...this describes me pretty well. I believe "rank amateur" comes closer.
They may be quick and dirty solutions to fulfill people's needs at the time. They may not be compatible with the newest versions of MediaWiki, or may use appallingly bad and even unsafe practices to get things done, which may lead to the inclusion of security vulnerabilities; commonly XSS and SQL injection.
For example, I'm pretty sure I used some appalling practices (like saving state by serializing data into the page output itself instead of to a session or cookie) in the Table Edit extension I just wrote...
But then I don't think it should actually be enabled on the Wikimedia cluster. Since extensions like these are being written by us amateurs to get things done on our own wikis, it would be nice to get feedback on which of our appalling practices to look out for. Perhaps someone with more expertise than the likes of me can add some more information to the sections on mediawiki.org about writing extensions and special pages.
But I'm NOT expecting the main devs to have the time to do that.
Jim
If we're to enable an extension on the Wikimedia cluster, then we have to be reasonably sure that we can trust the code. This means we have to review it for security and for sanity, and for performance. This takes time. Time which we don't have, on the whole.
Some of this can be skipped when the extension is written by an experienced and trusted developer and uses appropriate techniques to do things, although there is still a phase of review which (rightly) needs to take place, so it's still not instantaneous.
Another big problem is that people request the installation of the most ridiculous and pointless little extensions. There are vast arguments over how complex wiki text should be, and internally there is a ripple of discontent over how complex it has become. A lot of the community's thirst for nonsense was slaked in an efficient and effective manner by Tim Starling with the introduction of ParserFunctions. Unfortunately, now they want more. Variables. Arrays. All sorts of nonsense which, frankly, they *do not need* to write an encyclopaedia.
The community frequently bitches at the development and system administration teams to do things faster, but it is the development and system administration teams which have a responsibility to keep the sites alive, and which will take the flak - oh yes, there's a lot of flak to come from all our users - when the shit hits the fan, as it might do without the proper processes.
Rob Church
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org http://lists.wikimedia.org/mailman/listinfo/wikitech-l