1) uses MediaWiki syntax, which is infinitely better than C2Wiki, Wikidot, or most other Wiki syntaxes,
2) requires almost no installation; resides in a single self-editing file or resides in a pair of files (a script and a database), and
3) resides server-side, meaning that users wouldn't ordinarily be able to edit the Wiki.
I'm so used to documenting in MediaWiki format now that just about everything I write is in MediaWiki syntax. It bothers me because to display it to an end user, I have to: 1) parse the file manually and convert it to HTML, or 2) create a whole Wiki, which is silly when I'm writing just one page for a project.
If all else fails, I wonder if there's a simple parser which takes a file with MediaWiki markup and spits out a static .htm corresponding to it (e.g.,
<h2>for every ==,
<hr/>for every -----,
<b>for every ''', etc.).
I'm almost tempted to get back into finishing my DM-format MediaWiki parser. I previously abandoned the parser as "possible but unfeasible", which doesn't bode well for my motivation.