Comment by jchw
16 hours ago
Counterpoint: the problem was solved by an alternative, but was it solved without any tradeoffs? If the solution comes with no real tradeoffs, then that begs the question why React wouldn't simply adopt the solution... And I think this is not a theoretical nitpick, either: Angular actually has, to my knowledge, basically done this more than once. 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.
I think when you consider that the solutions sometimes may come with tradeoffs that cause problems that people like less than simply working around the first problem, it makes a lot more sense.
I don't think conferences, hype, corporate backing explain the continued success of React; they help but it's just not the full story. I don't think they explain the continued success of anything else, either; I find this to be a shallow and lazy dismissal in most cases. It's easy enough to find counter examples where none of these things, not conferences, hype, corporate backing, or whatever else you could think of, were enough to make something work. While these things definitely help keep something relevant, they can't do it alone.
In fact, in attempting to rationalize the success of React without acknowledging its strengths, I believe people have fundamentally reversed cause and effect. I think a lot of the continued success of React comes from people continuing to choose its set of tradeoffs over others even when they could tolerate risk. I think that conferences continue to be organized and books continue to be written because of its continued success.
The tech community on the Internet has largely matured enough to acknowledge why PHP was and is successful; yes, there are many objective issues with it even today, but it also has many strengths beyond just having existed for a long time, too. Do I think you should choose PHP for new projects, the way I might say for React? Well, no, not really, if I'm being honest. But, I think it was nice to see people grow up and acknowledge that actually, there were and are a lot of redeeming qualities to PHP, and what it did for a lot of people was pretty cool, fractals of bad design be damned. It'd be a shame to see this repeated again more for stuff like React and say, Go, just because it's not the dog in the race we wanted to see win. (And I would understand that position, because I fully understand why people like Svelte, or Rust. I just don't think there is a conspiracy, it's just that there isn't a free lunch here.)
> 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.
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.
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.