I thought some more about this and I see we are quickly entering bogus-land.
Chris & other "let's change everything" people: *what is the point* of spending a lot of time on migrating all the information to some different wiki-like system that uses a different syntax, has different features and different everything? There is nothing of substance to gain. We end up with same not very well organized information, but this time in a custom system that might not have all the features we had in MediaWiki and has a different a syntax. MediaWiki is used by thousands of projects and has a lot of people working on it, while your custom system would have only 1 user (inkscape.org) and 1 maintainer (you). We are not in the business of writing and maintaining wiki software. Furthermore, many more people are familiar with wikisyntax than reST.
Single sign on is the only real advantage your solution would have, but instead of removing MediaWiki you can just use one of the authentication plugins for MediaWiki. http://www.mediawiki.org/wiki/Extension:CASAuthentication http://www.mediawiki.org/wiki/Extension:AuthDrupal https://bitbucket.org/toml/django-mediawiki-authentication/src
You essentially want to spend a lot of time to make something possibly inferior. WHY? Forget about the wiki for now, it's not broken. In fact, it would work extremely well if it had a better skin (e.g. the default one!). Fix the main site first, then we can start talking about the wiki.
To sum up my points: 1. Don't try to fix things that work very well. 2. Standing on the shoulders of giants is good. 3. Fix the main site first.
Regards, Krzysztof