Move "Database search" into its own project

This was already discussed in the Search API BoF yesterday, and I think we pretty much reached a consensus on this (correct me if I'm wrong).

Currently, the database backend is broken, and I don't think I'll be able to fix all the outstanding bugs before the Search API 1.0 release. Releasing a stable 1.0 version with a broken backend, however, is of course far from optimal (even though it would of course be noted on the project page, and in the release notes). Having the backend in its own module, we would get the time to really fix all (or at least all major) bugs before doing a stable release for that backend, too.
It's also kinda weird that the (currently) best backend, Solr, is in its own project, while the Search API project contains a rather broken one. While it of course makes some sense from a requirements point-of-view (the DB backend works out of the box, you have to setup Solr and download the client to use the Solr backend), it also could backfire on the Search API if people try it out with the database backend, see bugs and errors and then assume that it's the fault of the Search API, not of the backend (as it's the perceived "standard" one).

The greatest solution would of course be if someone stepped up volunteering to (co-)maintain the database backend and attack some of those bugs. (It would of course have to be someone quite fearless of complex SQL queries.) I'd not even be against changing the whole structure in which data is indexed in the backend, if that made the whole thing more maintainable (but as far as possible without losing features or (too much) performance). In that case, moving it out into a different project would really be a profit. Without another maintainer, there would of course be the danger that it would go even less maintained than it is currently. However, if enough people are interested in a working database backend, they'd of course eventually have to step up and help with it in any case.

As much as I'd love to have a working database backend, there's so much other stuff to do, and the Solr backend is so much superior anyways, that it mostly doesn't feel to me to be worth the bother to attack these queries once again. But if enough people care about the database backend, I'm sure we could come up with a battle plan to get it working—inside or outside of the core project.

Anyways, request for comments on the future of the Database search module. As it is now, moving it to its own project is the most likely option.