Wikidot grew up around a developer-led process, which in the beginning made it dynamic and interesting. But over time, that process turned hostile to the Community, and led us to make really bad decisions, such as announcing fancy new features that no-one had asked for, while ignoring absolutely necessary (but boring) work that everyone was screaming for.
This was my fault, for being more interested in using Wikidot than fixing the process. It had worked, why mess with it? Well, the main reason that "new user interface" never landed, was that for three months we've been ripping apart the process and rebuilding it. We have a new internal issue tracker that holds everything we work on. We have a process that moves from request to design to prototype to development to testing to production. Obvious stuff, but we were taking shortcuts all over the place before.
There are about four issues in that tracker for easier registration: when joining a site, when making a comment, when trying to edit a page, etc. This is our highest priority work. When it's done, making the step from anonymous user to registered user, to site member should be smooth and compelling.
For view permissions, you have my vote (literally, I voted that request up ages ago). We've got two issues under development for that. One is private categories, which will be a simple admin function (category is private y/n, if private only site members can see its pages). The second is draft editing, which lets members edit a page and then publish it in one step.
For what it's worth, I cancelled the new home page. Maybe another day, when there is spare time, we'll start on that again. It was a good idea, but we do not lack in good ideas, just time to execute them.