Seems like we could use a CSS tag.
I never ever considered Bootstrap even with its incredible popularity and community support because of all the legacy bloat and also because I always used SASS instead of LESS for my projects (and I rather use plain CSS than have two CSS preprocessors in my toolchain).
So.. I really like this change of direction, really want to see how well it works out for them.
I’m not sure what you mean by bloat, but there’s been a SASS version of bootstrap for quite some time. It’s good that they’ve fully embraced it over LESS though.
I definitely chose the wrong term there by calling it “bloat”, but I was referring to all the compatibility stuff that added complexity to the entire framework.. that has been solved by CSS3 features that have been in evergreen browsers for a while.
One big example is the grid system all these frameworks brag about. It’s a big mess of float:left; width: $big_preprocessor_calculations_here; and it just shouldn’t have to exists at all.
Flexbox has been around for quite a while now, even though vendor-prefixed (but last I remember, CSS preprocessors rock when it comes to vendor prefixes).
float:left; width: $big_preprocessor_calculations_here;
I understand businesses (like the company I work for) value compatibility because of their customers.. and I’d be all right if Bootstrap decided to keep supporting IE8, but what is required to keep supporting it is something I would consider a reason to stay away from it.
It’s a choice, either you appeal to some people or to the others.
About the SASS port, I think I heard about it once but the thing is: when I have to choose between two frameworks/libraries, if using one of them means depending on both the maintainers of the main project and the maintainers of the specific unofficial version, I think I’ll probably need a very good reason to choose it over the other.
Interestingly, flexbox has support problems in every IE version through 10: http://caniuse.com/#search=flexbox
This means that choosing to use flexbox causes problems for a wider range of users on average than you would think (although, as you can see from the “Usage Relative” graph in that link, it’s still not a large portion of total browser usage).
w.r.t. legacy bloat – it seems like the amount of code that was removed in v4 that you’d classify as “legacy support” (e.g. IE <= 8 hacks) was pretty small. So it doesn’t seem like there was much of it to begin with.
It’s really disappointing to me that JQuery is still a prerequisite when all of the plugins are UMD. It seems kind of… lazy? That’s a bad word to use for open source projects but I definitely feel that it would be a great precedent to set saying “we don’t need to be loading all of this. use mixins. use UMD or some kind of js module system. make your websites light and well functioning” and this is where I really see bootstrap as a framework but in a way it can do a lot better than it is now.
Hot damn, semantic examples. I’m sold.