However, voting on development priorities is useful, as a means for the developers to get some idea what coin owners (somewhat an indicator of coin users) think is important. No one really wants to work on things that aren't wanted.
I see your point, but I mostly add the features that I want the most. If that happens to correspond with what lots of other people also want then that's nice.
I guess probably more people use the Qt client than the command line but I don't like working on it and so avoid it as much as possible. If lots of people want a better Qt client then I expect one of them will make it happen. If not, I guess it isn't really that important, and if they do that's cool with me - it doesn't affect the client I use, or the network as a whole.
You're right. A lot of open source gets done (especially when being done by volunteers) because the developer doing it wants the feature. Nothing wrong with that.