While Bootstrap v3 is largely stable, this is an interesting development given the effort needed to maintain a (soon-to-be) legacy version measured against the resources needed to work on the future version.
Will people hop onto v4 alpha and boost the development effort there? Or will they stand by stable v3?
Will people transition to other frameworks in the interim?
Will somebody fork/maintain v3?
Why is this a necessity in bootstrap, where other large projects have lumbered on without taking such an action?
Have other large projects succeeded with a similar strategy? Or will other large projects follow suit?
Perl 6, Python 3, rails 2, ruby 1.8, gtk, … Too many examples to list. It doesn’t always end poorly, but there’s usually a good three to five years of suck.
[Comment removed by author]
It is an unfair exaggeration to say that Angular 2 killed the project. There are a lot of angular projects being worked on or started every day. There are code bootcamps still teaching “The MEAN Stack”. There is even a three-day conference coming up this month.
However, I agree they will likely end up with a problem reminiscent of the Python 2 vs Python 3 divide. So many major changes were proposed just as React was beginning to steal Angular’s limelight among the Hype-Driven Development crowd, that it has shaken some people’s faith in the ecosystem. A sampling of those changes:
In discussions about Angular 2, I have heard people’s plans fall into four groups:
$OTHER_FRAMEWORKinstead.Where I would say this differs from Python’s situation is the existence of opinion group #2. I never came across anyone saying “Well, we already disliked Python, so rather than transition all our code to 3.0 we’re using it as an opportunity to rewrite in Ruby instead.”
Disclaimer: Though I am defending Angular as still alive, I’m just trying to be fair. I happen to fall into group #2, leaving rather than transitioning.
My first question is instead: Do people expect a CSS framework to go through such significant change that it needs significant resources to maintain?
Based on this situation, the answer is clearly yes, and my question shows a lack of perspective (lack of imagination?).
My perspective comes from usage at work, and from a few small side projects (half-dozen-ish) that use Bootstrap 3. However, in each case, I’ve only used one-quarter to one-third of Bootstrap’s total features. From that uninformed perspective, I don’t see it as a burden to be frozen on a set of stable, working code for the side projects. Most of my uses of it aren’t tracking the latest patch-release in the first place. Maybe not even the latest minor version.
At work we didn’t buy into any ecosystem of CSS or JS libraries that rely on Bootstrap. We also use rems anywhere possible, but Bootstrap 3 had to be retrofitted to use rems instead of pixels. That retrofit (née hackjob) predictably wound up being slightly brittle, so we’ll move forward to Bootstrap 4 sooner rather than later for the rems and flexbox support. Ironically not adding dependencies on tools built specifically on top of Bootstrap 3 will make it easier for us to transition to Bootstrap 4.
I don’t know if I’m going to bother moving to v4. If I decide to update my CSS, I’ll be more likely to just pare down and use what I really need, not use some giant bloated framework again. The grid alone is probably 90% of what we use bootstrap for.