I'm a little skeptical of voting systems. I do keep a close eye on the forum and the regularity features are mentioned are already a rough sort of voting system. The problems I see with formalising it are:
1) users vote for things which are technically an extraordinary amount of work, or are a completely different product, e.g. 3D. I'd hate to have to keep defending why we are not doing 3D when it's the most voted feature.
2) I am actually a little skeptical that people really know what they want, first with the WebGL shader effects feature, and then much more clearly with the multiplayer feature. Multiplayer was extremely highly voted for in previous polls. We developed it over several months - it was hard work, and involved quite a lot of research and prototyping before that. Despite the huge amount of heavy lifting the multiplayer engine does for you internally, networking is just a difficult topic in general, and I think that's why it's seen fairly low uptake. I think people see features in the voting stage as magical solutions that will solve everything with no effort, but realistically nothing ever turns out like that. I think we documented it well, it's just hard to design multiplayer games. Maybe it will see increased uptake over time, but so far it was an awful lot of effort for a feature that has seen niche use in practice.
3) Users don't see the internal technical complexity. For example a common request is to get the editor tabs to persist the same colors between launches. Seems simple and trivial, but I've looked and it's buried somewhere in the black box of our pre-compiled UI library, so it's really difficult. So then you might have an apparently easy feature really high voted, and everyone's wondering why don't we just do it?!? Accusations follow that we ignore our users, etc.
4) Lots of suggestions come without compelling use cases, e.g. "why don't you make a feature to return the Nth prime number?" Well... why? Who needs it? What do they need it for? Why is it so important? Lots of one-line suggestions lack a thorough definition that identifies it as something really compelling that deserves attention, or are so broadly defined or vague that it's hard to know how to actually address it with code. Every feature we work on comes with opportunity cost - the possibility we're not working on something else that could be even more important.
Obviously we wouldn't exist without our customers, and exist to serve them. However there is a careful balancing act to be made and I think user voting is a pretty distorted version of that. Also, I'm pretty sure Tom has no time to work on it