I've been using Go and contributing a little to the project. It's a cool language and the guys who review your code (Rob Pike, Russ Cox) really know their shit.
Almost nowhere. It hasn't even been adopted within Google. But every language starts out in this state. It's impossible to tell where it'll be in five or ten years.
I wonder what Google is planning to do with this? What do they plan to offer with Go that is not already available with other languages? Does it have any connection to their recent admission that they might never be able to make python as fast as C or fix some of it's limitations? Maybe they are a trying to do with Go what they did with Chrome (for browsers). They certainly have the engineering the programing muscle to make it happen.
I don't know if they have anything specific planned. Google hired a lot of former Bell Labs people, several of whom are working on Go. They are probably letting those people working on new ideas as they wish, with the hope that people who made stuff like Unix will make something else game-changing and worthwhile for the rest of Google eventually. Sort of like MSR, except not just academic research.
It does make sense. What do you do when you have a surplus of extremely talented programmers and engineers? Give them the freedom to do anything they want. I think Google's 20% rule is an excellent idea.
Go began as a 20% project, and grew into a full-time project. There are now a handful of Google engineers working on Go. It is not part of some over-arching corporate strategy (that I'm aware of), except tangentially in that Google seeks to improve the software development ecosystem globally. From the engineering resources given to the project, it would seem Google is committed to making it a success.
It grew out of scratching some itches of C++ (cumbersome and error-prone concurrency, unnecessary complexity, monstrous build times). The interest from the dynamic language camp has been something of a surprise, although not an unpleasant one.
Go scratches so many itches from other systems programming languages while remaining nearly as fast. It makes code cleaner and easier to manage. I say this as a college student working on a 7000+ line side project in Go.
Garbage collection, concurrency primitives, static duck typing, general rational design. So worth it.
As for Google's strategy, well, it keeps the Bell Labs guys working on good stuff while giving Google a good new language to write their low level systems in.
They're still considering it, but nothing firm yet. Like Java, it's usable without generics by using interface{} when you would want a generic type. But admittedly it'd be nice to have for some things like new containers.
I really like the new development of goinstall - a package manager for Go. I especially like the choice of native support for GitHub, BitBucket and Google Code.
