Brion:
If the project ever actually makes it as far as having some software, of course it would be publicly tested first. So far as I know no code has been written
I'm actively working on it, though of course I'm not going to commit something incomplete to CVS. It will still be a while until we get from where we are to a working implementation (I'll post an updated timetable on Meta soon). This is a very large project and I want to get it right instead of hacking together some simplistic template-based solution over a weekend. It needs to be flexible and scalable beyond UW itself.
So, personally, I think a discussion about the usefulness of UW is at this point a bit premature, as the development has been funded and is underway, and the decision about whether we want to do this should be made when people have had a chance to look at a working prototype and comment on it.
There will be no automatism; just because someone develops something, it's not necessarily going to be used or even accepted into MediaWiki itself. Gerard's article is a solid presentation of his view on where we are going to take this, which I share, but it should be self-evident that this can only happen with the approval and support of the community.
Best,
Erik