I have been using Ember for over 2 years, since v1.0.0-pre.2, and it hasn’t been easy.
We do not use ember-cli. We use Visual Studio as the IDE and have built a handlebars pre-compiler and a bundling mechanism to deal with what we need since these things were not available in the pre2 days.
We have a rather large project that is a front-end to an integration back-end. Around 300 people on the project at any one time but that number includes developers, testers, and BAs divided up into quite a number of scrum teams.
Our first major hurdle was upgrading to v1.0.0 where it took us probably upwards of a month to get our code stable again.
We are now on v1.6.1 but an upgrade would require us to re-write all our controls into components. Components didn’t exist when we started and singletons were all the rage in the Ember world
However, for an upgrade we’ll probably need a month or two just to get our framework back up-and-running and it would then require changes to all of our 300+ handlebars files. This would also require full regression and re-deployment. Our aim is to do this after our largest go-live scheduled for around the middle of July. My estimate is that we’ll only be stable again after around six months.
Seeing as how this system is used to manage around 8,000,000 customers (and that number could skyrocket) it is rather crucial that stability is maintained.
If you do not have these types of constraints it will be easier to stay up-to-date. We can’t.
So do I mind using Ember? No. Would I use Ember on any projects that I have a say in? Probably not.
That being said changes in some other popular framework(s) would also irk me. Apparently Angular v1 to v2 is somewhat of a challenge.