On 29 May 2009, at 14:50, Daniel Silverstone wrote:
Option 2: Mediawiki
Mediawiki is a very popular and well-known wiki engine. Used by
Wikipedia and various other sites, it is richly featured but unlike
trac, it is just a wiki engine.
Pros: Only a wiki engine, so we're free to implement the other
in other ways without confusion. Already has an example instance on
machine which currently runs our SVN.
Cons: Only a wiki engine. Written in PHP which means we cannot ever
consolidate its functions onto pepperfish (where the rest of the
websites are). Is complex and backs onto MySQL which is yet another
I would warn that I've dealt (briefly) with the MediaWiki code
before, and it /is/ pretty horrendous, as one may expect for a
project that has "grown" in PHP. Unfortunately, though, this is a
pretty common flaw in wiki software. It will almost certainly go
straight to the top of top(1) on your server, and this just gets
worse if you want to borrow templates from Wikipedia, as you'll find
yourself needing the option that runs all output through HTMLTidy. I
really wish I was making that up.
On the flipside, this is balanced by its dominance meaning that there
being a lot of people applying gaffer tape to it to keep it
progressing and vaguely working. There are also extensions which may
be useful for dealing with bug reports etc., such as Semantic
MediaWiki and Semantic Forms; piles upon piles of ugly hacks,
but the user experience can be that you can then create instances of
types of things, and have a nice little form for people to fill in
most useful, as it's the sandbox.
The only live site I know of using this is Chickipedia, which was
mentioned at the WWW2008 conference. Probably not advisable to visit
| Philip Boulain PhD student | --+--- | Modesty helps one to go for- |
| IAM, ECS, Uni of Southampton | |@.... | ward, whereas conceit makes |
| |..f.. | one lag behind.-Mao Tse-tung |