Running with my own tangent, IMO Groovy went about this backwards, where each IDE grafted their own type inference on top of stock Groovy in their own unique way, instead of just specing it as part of the language.
Granted, that was what they had to do, because it seems at the time Groovy's committers were too proud to realize that their users actually did want static typing. That seems to be changing with Groovy 2.0.
Anyway, sorry for the grumpy comment on your post. Auto completion is wonderful and amazing, it's awesome to see people building IDE-ish features like this in lightweight editors. I'm always psyched to see that.
I probably cba to add [String] annotations.