Well, SMW actually competes regularly with turnkey applications already, and sometimes wins; that's partly because it has features that no one else has (version history, I would guess, is among them). It has its advantages, and disadvantages, like every application ever written. You could make the case for some feature or another on its merits, but arguing that SMW *needs* feature X in order to compete just doesn't seem fruitful.
John - I have nothing to do with the SMW query language, so I assume that wasn't directed at me. But to clarify what I was saying before: there are certainly features in Semantic Forms that get requested frequently, and that it would be great to have - dependent fields (like one dropdown based on the value of another), which you mentioned, is one notable example. But there are other features that are less useful because the functionality can already be done another way, or they're too hard to implement to justify the effort, or both. It's on a case-by-case basis, obviously, and hearing the specifics of the data need goes a long way toward determining which it is.
Alex - a development roadmap for each extension might be nice to have - I'd definitely like to see one for SMW itself. :) I have my own private to-do list for SF and my other extensions. I used to send out a "planned features" email to the SF mailing list, back when there was one, every six months or so - maybe I should start doing that again.
Let me take a moment, though, to plug the "Funding needed" page - I don't think it's been publicized officially yet:
This is as close to a public roadmap as there is right now. My project on there, "Semantic Classes", is in fact my highest priority at the moment - I hope to start working on it as soon as I can get funding for it.