Gregory Szorc's blog

Gripes About MediaWiki

In trying to figure out what I want to do for MediaWiki as part of Google's 2006 Summer of Code, I am listing some of my gripes about MediaWiki, the software that powers Wikipedia and the Case Wiki.

User-visible

• No WYSIWYG. Unfortunately, not ready for prime-time.
• Discussion method is horrible. Why not just integrate a tailored discussion board into MediaWiki. This is a front-runner.
• Lack of robust permissions system. Can you protect specific articles? Articles with a prefix? Limit rights for specific sets of articles to specific groups? No. This is another front-runner.

Back-end

• PHP 4. PHP 5 is the way. Not likely to be a viable project.
• Use of global variables. Seriously, pass them as function parameters or have an object registry. This is on the table.
• Configuration architecture. No web-config. No ability to run multiple wikis from one installation. This is also on the table.
• No web services. Off the table according to developers.

You can ping this entry by using http://blog.case.edu/gps10/mt-tb.cgi/7776 .

Make MediaWiki use HTTP codes correctly.

I swore I saw them commit that in the code sometime.

http://bugzilla.wikimedia.org/show_bug.cgi?id=2585

I got tired waiting, so I patched the Case Wiki to issue 404's. Any other status codes you want?

201, 300, and 410 would be rockin'. 401 would be nice, too, but doesn't jive with CAS.

So...

201 on edit completion
300 for ???
410 for deleted pages

I'm weary of 401 because I don't see how it integrates with MediaWiki.

201 for creation. Edits should just issue normal 200.

300 for the disambiguation pages... ??? ... hmmm... okay, skip the 300 thing. It's hard to map the semantics.

Actually, the more I think about this, it's sort of pointless. In the future, wikis will be editable over the APP which should make everything wonderful and is entirely more useful than trying to hack correct HTTP codes into MediaWiki.

Greg,

I think that while the forum/discussion on MediaWiki is the most intuitive, needed, and useful addition you could possibly make, I also firmly believe that it would be the least likely project that Google would support.

I think they really like the idea of unique and interesting research coding opportunities that you could present. A forum is far from unique and definitely not of grant-level interest. Just my opinion. I really hope you write a proposal that rocks them and nets you an interesting summer project.