Kristen’s classification of the four eras of JavaScript frameworks feels intuitively correct. A further simplification is essentially: 1. jQuery 2. Backbone 3. React 4. Next.js. Those were the big names from the eras and the similar technologies feel like obvious siblings. It’s easy to point at shortcomings, but overall I feel similarly positive:

Overall, I think the JavaScript community is heading in the right direction. We are finally developing mature solutions that can build full apps from the ground up, solutions that are not “just a view-layer”. We’re finally starting to compete on the same playing field as SDKs for native apps, providing a full toolkit out of the box.

Nicklas Envall’s history paints a similar picture, as does Josh Justice’s Frontend Web Architectures (I enjoy Josh’s naming of the first era: “JavaScript Sprinkles”). They both beg the question: what’s next? It’s too early to know. I think we’re still pretty firmly in the meta frameworks era. We see attempts at newly-coined terms and development paradigms, like Astro with “islands” and Qwik with “resumability“, but they don’t seem to be quite worthy of declaring a new era.

Perhaps more era-worthy is all the notable work is being done on JavaScript runtimes themselves (so many!), the compilers that run on them (so many!), and the steps we see toward the server.

Nicklas asked are compilers the new frameworks? Maybe?? Personally I’m fascinated by the foundational tooling work being done. People are seriously hyped on Vite, for example. Vite isn’t a framework, it’s just a build tool. It might help power the framework you like. It’s like being stoked about your carburetor and not your car. That’s worthy of consideration.

Lower level than that, esbuild is just a bundler and people are still hyped on it. Investment is happening as well, both literally with things like Void(0) and with companies internally investing like we see with Turbopack.

As ever, you don’t need to care. I like to think we’re getting more sophisticated on choosing when we need to care. Certain holotypes of website don’t need a framework at all, and avoiding one brings a slew of advantages from performance to long term maintenance ease. There is palpable pushback as frameworks get harder to use. No build system is also possible, thanks to neat modern tools that don’t need it and platform improvements. There is a cost to (short term) convenience, they say.