Let me start by saying that MediaWiki is great wiki software and I have enjoyed
writing many extensions and plugins for it. The requirements of Wikipedia have
demanded features that are hard to come by in other wiki software, especially in
the areas of caching and optimization. However, this is a double-edged sword,
as development decisions in MediaWiki are often filtered through the question of
whether they are desired for Wikipedia. If this litmus test is not met,
features are abandoned and never see the light of day. If so, they are
implemented to the delight of Wikipedians. I don't mean to criticize this
methodology and I certainly don't wish to bad-mouth MediaWiki (after all,
someone was kind enough to make the product open source and kind enough to make
it generic that anyone could use), but I have a few questions that have been on
my mind.
My questions to the MediaWiki developers are as follows:
Will features perhaps never desirable for Wikipedia ever be considered for
MediaWiki? For example, wiki spaces (much like XWiki or Confluence) that either
exist as self-contained wikis or as areas that are more stand-alone (e.g.
attaching documents to individual pages). Another desirable feature is a
fine-grained permission system, although I argue this goes hand-in-hand with the
aforementioned concept of wiki spaces.
Will MediaWiki development ever take a significant step away from Wikipedia
so drastic overhauls to the core can be made? Right now, MediaWiki development
is mainly implementing features on a production site and spewing them back out
in the form of a new version release. Will this model ever change?
I see MediaWiki as much more than the software that powers Wikipedia. One
only has to look at the hundreds of independent wikis in existence that are
powered by MediaWiki. However, there is much ground that can be won. I argue
that because of Wikipedia's popularity, the MediaWiki wiki syntax is the most
recognized of any wiki syntax in existence. Combined with its proved track
record, MediaWiki is immediately a front-runner for anyone wishing to deploy a
wiki. However, we all know MediaWiki isn't ideal for some situations. Well,
basically it is ideal for only one simple use case: a wiki focusing on one
subject matter with blanket access control to all content therein. Any other
use is a stretch. In other words, there is a considerable market share in which
MediaWiki can't even compete.
MediaWiki is great software and will continue to be great software. In order
to take it to the next level, a slight departure from the Wikipedia-centric
development cycle will be required.
I kindly ask that the MediaWiki core developers lend their thoughts on
this matter.
Gregory Szorc
gregory.szorc(a)case.edu