Regarding Closure and its libraries, my impressions are that it gives tremendous maintainability and structure when you have a large team and a long-term project, but there's still a lot of boilerplate when you need things like a simple data binding. Having searched around, I've found that Angular (1.2) provides a good balance between structured, easy-to-find-things code, and fewer keystrokes to iterate fast on features. I wish React were like that, but the lack of best practices in the Flux architecture makes it hard to know where to put things - I found I was constantly worrying about shooting myself in the foot when I made architectural decisions in React.
I do think that Angular's going to get a Python-like version split, and it's refreshing that plenty of people are still writing Python 2 libraries... because I'm not leaving Angular 1.2 any time soon!
Keeping the various spellings of the name right is a challenge, but it seems to work very well. Yet it is so far from the rest of the ecosystem. When you roam across some interesting JavaScript library online, there is basically a 0% chance that it does things the Closure way.
But it lacks a lot of magic. It's based around reusability, maintainability, and performance, and not just features. Like you said, features get blogged, maintainability gets (silently) used.
Well git isnt a javascript compiler and git doesnt need the JVM to run.
In client-side land, it's like there's inverse relationship between "blog chatter" and "real-world use in large projects". The tools that get all the attention are only used for personal tinkering, or by startups that mostly won't exist in a few years. Over the past several years, if something seems popular then that's a sign it's probably a children's toy for which you shouldn't stick your neck out within your company.