While I certainly agree that the code is currently Toady's to do with as he pleases, I do honestly believe that an open source approach would grant better results, faster.
Look at all the tools people have been creating, for free, in their own time, to interact with DF in different ways and to fix the issues they found with the game. Imagine what we could have if the guy behind Stonesense and Toady would be able to work together on a single code base...
The most difficult obstacle for an open-source project to overcome, the one thing that makes or breaks success, is a community. And DF already has one! And a pretty good and tight one at that, because we're all here because we like DF and that's not such an easy thing to do, all things considered. So we have quite a bit of an 'in-crowd' thing going.
Sure, small outlying elements may try to 'steal' the code. Good luck to them, survival rate for forks is what, 1 in a thousand? After all, such a fork can only survive if it creates its own community, but for that it has to fish in the exact same pool as its parent and all the fish in there will already know who's the original and who's the wannabe. Khazad hasn't seen any development in 235 days now, looks dead to me.
And you don't even have to open source everything. Toady could very well keep the geological engine separate and closed source. Heck, that sounds to me like the sort of software a savvy developper could make money with, licensing it to just a handful of larger companies. The market for specialised software is very different than the one populated by whiny gamer fanbois.