Issue #25: Bring Me To Life

🎶 Frozen inside without your touch, without your love… 🎶

Time flies, new metaframework releases come and go while we’re trainspotting the cars of this never-ending chain of innovation, technological excellence, and (some) drama. Who’ll hit the ground and who’ll get high this time? Let’s see…

The Good

The pair that will probably never fall apart now came up with new awesome releases each. (Which probably hints on some common release schedule in future but… nah, who’d do that…)

Next.js’ had announced version 15.4 (weirdly enough, I wasn’t able to find the actual 15.4.0 release page on the GitHub, only the patches and canaries, and npm says “This version was published in error. Please use next@latest” which is too suspicious… What have we missed there?! Failed Nuxt merge attempt or something?! 🤔) bringing (“almost”) full Turbopack compatibility and (“some”) stability improvements. In fairness, the announcement mostly emphasizes the updates coming in the future versions (similar to my daily standup reports) so it’s hard to put any accents here, but as with any tools providing maintenance mode for thousands of enterprise teams, the improvements are rarely very visible but at the same time pretty important and anticipated, at least for some of the customers. There are some neat DX improvements too making our work of building and fixing Next.js applications a little bit easier.

On this background, the update from Nuxt looks much more significant. But still…

This is a stability-focused major release, introducing a few thoughtful breaking changes in order to improve development experience.

So some breaking but no groundbreaking changes (which is good!) are awaiting in this v4 update, the largest and the most notable of which is the changed project structure. Probably not too much for the first major release since the previous one saw the light 2.5 years ago, but that’s something we love Nuxt for — stability and persistence (you see — the guys don’t declare “stability improvements” — they just don’t need them 😉). Even the project structure changes are optional:

Don’t want to migrate? That’s totally fine! Nuxt will detect your existing structure and keep working exactly as before.

And even deployment platform vendors outside the Vercel realm (like Netlify) also support this consistency with easier setup provisions, which is beneficial for everyone.

With that, both Next and Nuxt become go-to choices for folks valuing stability and predictability over unconfirmed innovations and false hype-driven ideas (or even just benevolent fantasies of some framework authors).

The Bad

And we know some negative examples of that right? Abstracting away from your personal attitude to this or that tool or vendor, there were notable stories about breaking changes in the JavaScript [meta]frameworks history. There were pains of Angular 1->2 migration many years ago (I still try to erase the v1.3 forever tattoo from my chest) and less distant scandals of Svelte 4->5 rebuilding (because of which I still cannot psychologically switch from 4 to 5 sets of my morning push-ups reps), to name a few.

In the recent times such confusion was greatly associated with Remix / React Router evolution, where the former transitioned to the latter and vice versa several times. Now the ways of these Rs seem to diverge quite a lot, with Remix becoming something different, and React Router slowly going away from the framework mode lately and starting to couple more tightly with the latest and greatest RSC stories (even more confusingly declaring that in the Remix blog 😵‍💫).

Now there’s a savior who had connected the dots (at least for me) on the Podrocket podcast episode dedicated to Remix the last week! Brooks Lybrand had explained where both tools are heading to and why — at least on the level that is possible in all the current hassle in the framework development process now. It’s confusing, it’s misleading, but also it clears up a lot and there’s some hope that the two tools will form up into something definitely decoupled and hopefully meaningful in the near future.

The Noteworthy

It’s good that we have some meaningfulness anchors right now (otherwise I’d personally go mad curating this newsletter 😅). For instance Astro and their new minor release 5.12 bringing TOML support for content loaders (love it — this format is just sane, as compared to many others!), local platform capabilities for Netlify (narrow focus, but that’s probably something you should try too!), and some other nice enhancements and bug fixes.

With that, I’m really glad there’s such a pluralism still in the country of metaframeworks and everyone can find something matching their mental models and technical preferences. The more, the better, just don’t forget: no cargo cult and no JSX.

Just kidding, I am personally fine with the cargo cult.

👋