Then you get aggressive opinionation about the projects, but really that's a secondary phenomenon.
One problem is there are no objective metrics for framework or language quality. I don't think anyone knows if objective metrics are even possible, never mind how they would work.
So you pretty much just make a choice that works for you and stay with it until something obviously much better comes along.
And if it doesn't - if it's obviously much different, but not so obviously better - you can still be getting useful work done.
The seductive promise is that Framework X will make the job take half as long and produce half the bugs.
I doubt that's ever true in practice, for general values of X, especially when you consider learning/retooling time.
Actually, the main reason I was feeling like I might want to switch from Grunt to Gulp was the extreme and measurable time difference between running the two. Honestly the "it's easier to set up!" arguments fall somewhat flat; it's not HARD to set up Grunt, and you don't need to do it often, so that stinks of "premature optimization." Broccoli seems to entirely be about how clean and tiny the configuration files are; not interesting, because that means it's actually harder to understand how to change them to do something the designers didn't expect. If it's even possible.
It does feel like each group has their own little fiefdom and they want you to use THEIR tool rather than a competitor. The Grunt main page says "THE JAVASCRIPT TASK RUNNER", as if there are no other options. As a new JavaScript developer, it made me think that was just the only option.
I am much more impressed with the integrity of open source projects that actively advertise the options -- especially when they cite advantages and disadvantages of the options. Actively misdirecting you to believe that there no options is pretty much the opposite of that philosophy.
So the field is a firehose of high-profile projects, whether or not they're needed.
It's possible a very large and complex generalisation of Conway's Law, between groups that barely communicate:
http://en.wikipedia.org/wiki/Conway%27s_law
Then you get aggressive opinionation about the projects, but really that's a secondary phenomenon.
One problem is there are no objective metrics for framework or language quality. I don't think anyone knows if objective metrics are even possible, never mind how they would work.
So you pretty much just make a choice that works for you and stay with it until something obviously much better comes along.
And if it doesn't - if it's obviously much different, but not so obviously better - you can still be getting useful work done.
The seductive promise is that Framework X will make the job take half as long and produce half the bugs.
I doubt that's ever true in practice, for general values of X, especially when you consider learning/retooling time.