upvote
> And React certainly doesn't seem afraid to make big or breaking changes, from fibers to suspense to hooks to context, so I don't think it is resistance to change at play here.

As far as I know React hasn't had any such breaking changes. Those were all added on top without removing anything. Class components without hooks still work fine, for example.

reply
As a very long time React user, React.createClass is certainly gone. So are some old versions of the context API. Although it never broke any of the guaranteed behavior, Fibers broke a lot of behaviors that were not guaranteed. Upgrading the major React version in any very large application is likely to cause at least some issue to shake out.

I think the more correct thing to say, at least in my opinion, is that React never breaks functionality without good reason. If you are following the React best practices and not relying on unintentional implementation details, using hooks/lifecycles/APIs correctly, your application will not break for a very long time.

If there was a sufficiently good innovation that was a pure win with no downsides, I believe the React devs would do their best to integrate it in a minimally disruptive way... But just like the adoption of ES classes, I do expect that eventually they will drop support for "the old way" when the time comes.

reply
I think we agree, though I can see why it might not have seemed that way, since I didn't draw attention that users might rightly conclude the overall tradeoff is worth it, even if another tech is sounder in some minor aspect.

React can adapt, and though Web Components might still have some claim on X, Y, Z, React still is heavily preferred because all the other things are so much more important to developers.

For what genuine gaps remain, there's lots of React users that can help others through it.

reply