In short, here is what yeoman currently claims to be:
Yeoman is a robust and opinionated client-side stack, comprising tools and frameworks that can help developers quickly build beautiful web applications. We take care of providing everything needed to get started without any of the normal headaches associated with a manual setup.
And here is what it would be after the direction change:
Color me unimpressed.
For me, what was most appealing was that it was a single tool unifying a whole bunch of pieces (e.g., grunt, bower) into one official workflow. Not that it had scaffolding.
So now I'm looking into brunch again.
One of the reasons I picked yeoman in the first place, was frankly just the speed with which it had taken off. I didn't know enough yet to really judge the projects on their technical merits, so I used popularity as a marker.
If you're looking for a yeoman vs. brunch discussion, here is one from someone who would know a thing or two about the differences.
So far, the difference that sticks out for me is that, in brunch, it
looks like you don't need to manually add script references any time you
add a new library. Simple thing? Probably. But it confused me that
yeoman had this elegant way of adding new things
whatever, yet that it didn't somehow magically wire it up for me as
well. brunch appears to take care of this for you, so long as you drop
the scripts in an appropriate folder.
In some ways, I think I'm kinda looking for a solution to a problem I don't really have. The problem I have is mostly mental. I want building a new test app to have as few steps as possible. I want to be able to jump right into what makes my app different rather than all of the boilerplate code and steps currently required to get something going.
I want a built-in webserver and build process so that I can be sure that filepaths that worked in development won't suddenly break when it gets pushed into production (however, given that I'll probably ultimately be integrating things with middleman, this particular problem can't really be guaranteed).
I think what I possibly actually want is a better coffeescript repl. I want the fire-it-up and just play that's so pleasant with pry. Just tonight, I was trying to work through some logic I hadn't quite sorted out (and as of this writing still hadn't sorted out) and I was actually using the "Try CoffeeScript" thing on the CS site, but then I realized I needed underscore or sugarjs for some functionality and I had no clue how to add it.
This sort of thing would probably be a perfect use case for jsfiddle, but it's like, I want the freedom to play. To commit syntax errors on a single line and immediately know where the problem is. But CoffeeScript's REPL still feels extremely primitive at this point (poor multiline support and either no readline support or it's just that mine ain't working). In short, it's no fun to play with.
But getting back to the topic at hand, given the community response to the proposed change in direction on yeoman, I kinda feel like the shift is going to be inevitable. However, I will be interested to monitor what actually happens.
I mean, especially going back to what I quote from the web page and what I remember seeing in the screencast. It just seems like a strange direction to take when they've never really made the scaffolding out to be some must-have feature.
And for me, at this point, it's not.