MediaWiki stores in MySQL on our system, but I bet having the ability to do a text dump of every article is something that could potentially come in handy if we do decide to go with a different engine, and I'd wager Perl would be good at it too. Thanks!