>>Because it is an extremely effective model for building UIs.
Is it? Have you ever tried? I think it is extremely hard to build reponsive UIs that work on every platform and there are no UX problems. We have to deal with a giant amount of boilerplate in many cases because using vanilla things (JS especially) won't cut it anymore. I am not even sure how many frameworks are targeting this space and try to solve the typical issues with these technologies and fail miserably. I think some of these issues are finally started to be handled the right way (Elm or ReasonML) but there are still some painpoints left (CSS). The question still, is it worth it? Couldn't we do better? The performance implications (energy consumption) for an average website is just hilarious. We should have a day where everybody disables JS and measure how much percentage of our computing capacity is going to render websites because we do not want to move away from HTML which was not (by any strectch of imagination) designed for this to start with. If anything, this should be a concern to most of us.
> I think it is extremely hard to build reponsive UIs that work on every platform and there are no UX problems
It is hard on native platforms as well. Developing for mobile is not that far off from web today, everything converging towards flexbox or constraint solving, reducers etc.
I think you’re conflating the problem of a bloated JS stack with the platform itself. We were building perfectly fine interactive UI without any of this crap ten years ago, and can still do it.
HTTP and HTML were not designed to do this. It was specifically static and stateless. We are retrofitting statefulness and dynamic content to this platform. Imagine if we designed a stateful and dynamic platform from the ground up that could be implemented with minimal effort to every platform and started to use the right abstractions with the right set of tools.
Is it? Have you ever tried? I think it is extremely hard to build reponsive UIs that work on every platform and there are no UX problems. We have to deal with a giant amount of boilerplate in many cases because using vanilla things (JS especially) won't cut it anymore. I am not even sure how many frameworks are targeting this space and try to solve the typical issues with these technologies and fail miserably. I think some of these issues are finally started to be handled the right way (Elm or ReasonML) but there are still some painpoints left (CSS). The question still, is it worth it? Couldn't we do better? The performance implications (energy consumption) for an average website is just hilarious. We should have a day where everybody disables JS and measure how much percentage of our computing capacity is going to render websites because we do not want to move away from HTML which was not (by any strectch of imagination) designed for this to start with. If anything, this should be a concern to most of us.