Frankly, as Wikidot gets larger, I think it will do so mainly with "read-mostly" users. They will be drawn to specific sites, and will mostly likely actually edit pages once in a blue moon. Personally, that's even me with respect to any site I don't own. And that's me over at Wikipedia. I've probably edited no more than a dozen times for some minor change.
I don't think you'll get much or any interest in this class of users in community. You'll probably be hard-pressed to even get them to register since many will have no need and/or their site does not allow/encourage editing.
And that may be where the technical has to cross the social. I guess I look at Wikipedia, and how they've had to struggle with open editing and vandalism/edit wars. I think there needs to be more tools around being able to "manage" a site's own community, particularly since sites here don't often have the manpower to keep an eye on things from minute-to-minute, and with that the case, most probably restrict editing rights to a trusted circle. It would be interesting to know the demographics of sites here: how many are "mostly open" and how many are "mostly closed"?
Tools that I would look at are ways to moderate edits, as well as being able to develop changes in private (even among a group a group) before deploying. In other words, use principles of source code management (e.g. Subversion, Git, Mercurial) to wiki editing.
And the final part of the social aspects is what I've mentioned before (and there's a page on weneed somewhere), where I can use a single user login, but present an identity of my choosing to a site. In these days of people needing to separate their private/professional lives to avoid the "Facebook" effect, it's a real pain to have to try to do this with separate logins (and very error-prone and irreversible if you screw up). Since my Wikidot usage began with one of my hobbies, that's all I use it for. I've thought of using it for blogging, or more professional use, but the identity issue throws cold water on these ideas.