Issue #20: Omen
The news in the tech ecosystem is built on reflections from the past and reminiscence of the future. Sometimes it makes a lot of good sense, but sometimes there’s nothing except for poor naked marketing and attempts to hide sales pitches behind the fig leaf of developer relations. The last week was full of both, so let’s try to separate the wheat from the chaff and dig through the most interesting stuff as usual.
The Good
A bit ahead of me releasing this Issue #20 of the Metaframeworks Weekly, the Angular team has released their version #20 of this one of the most popular frontend frameworks which, starting from version 19, transformed into a kind of metaframework too, with its isomorphic SSR-driven capabilities (not even mentioning AnalogJS, bringing Angular to a whole other level). I have a long-lasting Angular-related Stockholm syndrome case, working with this tool from its first version (when it was even called differently). As with several latest versions (after v16), this new one doesn’t bring any groundbreaking changes, rather it finally improves what was started before and just begins what was planned after these previous improvements… In short, there’s a good trend of Angular becoming a solid, performant way to build complex web applications, though for quite a while it brings a lot of grunt intermediary migration work for developers using it.
The other tool loved by Angular developers (and not only by them, of course), Storybook, incremented its version to the long-anticipated (and tested) number 9, with some significant updates I had already talked about here before.
The lean, mean component testing machine
The component testing capabilities of Storybook in this version come to an all-encompassing state indeed, allowing you to use it as a whole testing pyramid for the front end of your web application (where tools like that can replace all the alternative approaches less suitable for the context of a user interface). I truly believe the Storybook team is one of the most productive and collaborative (in terms of the part of the ecosystem they integrate with) teams in the dev tools space, which is moving faster and faster in the right direction.
The Bad
The Redwood team also persists in moving in their chosen direction and proceeds with their storytelling about the philosophical role of personal software in modern web development, and the place of Redwood SDK in it. But instead of making something scrappy like some passionate folks do, they go down some weird side tracks. This time it’s also about the place of Cloudflare in it, which is funny, because we see
No matter what choice you make, it’s always going to be vendor-locked in.
fighting with
Platforms like Cloudflare or Supabase shine.
And even if
You avoid paying these [vendor lock in] taxes repeatedly.
still
Switching out something, even if it’s open source and self-hosted, means that you’re rewriting a lot of code.
Yes, we have a lot of tools in the metaframeworks ecosystem (and in the JS ecosystem in general), but maybe for some custom needs, these tools are ideally laser-focused to solve their task. Don’t trust anyone. Choose your way wisely. Even (and especially) with personal software, where your technical decisions are absolutely your own convenience and choice.
And if you think that the JS ecosystem overwhelms you with its too comforting choices and lock-ins, check out other alternatives for building web projects and the “pleasant” experiences people have with them. Maybe React is not that bad, huh…
The only thing that bothers me in today’s JS ecosystem’s dynamics is the vendors’ inclination toward loud alphas and betas. Don’t get me wrong, I remember the book that partially created our generation of web developers, and I am an impatient developer myself. But in the world of broken stable software, marketing half-finished tooling versions is probably too salesy and haunting. I understand when it’s made by hype-prone React bros (see the previous issue), but when one of my favorite tools does that (looking at you, Vite!), I get a bit disappointed. Again, there’s venture capital behind that now and the guys need to deliver loudly, and get their word out through different new partnerships and integrations, but in the FOSS software world, it’s a bit not comme il faut. Even the aforementioned Storybook, which is also near and dear to my heart, and which had loudly released the beta several newsletter issues back, went this route. Same for Tailwind, Zod, and many others (Nuxt guys have cried about their alpha recently too!). I don’t think it’s a good trend, neither in software development nor in software marketing.
The Noteworthy
What can be considered a good trend in metaframeworks tooling development and marketing is the commoditization of some underlying techniques and approaches and coming up with good best practices around them. It both justifies the corresponding technological choices and helps to find the most efficient solution for a local problem. One of such things in the metaframeworks world is server functions, which are the underlying mechanism for client-server communication for lots of the tools. TanStack’s Jack Herrington talked about server functions on the PodRocket podcast last week, and this conversation and insights raised there were very approachable and (I’d say) somehow uniting for the ecosystem. Of course, TanStack’s approach to them was shed in better light there, but that’s not the point: the point is that at least in this direction, metaframeworks vendors are more or less on the same page.
And one of the server functions technology consumers, Astro, released its new version 5.9 with thoughtful (even though experimental) security improvements and lots of minor shenanigans we always can expect from the team. No BS, no fuzz, just stable and continuous delivery. That’s what one should expect from a good metaframework tool of one’s choice (not Astro specifically, there are others too).
These last news bring some anti-hype hope into this week’s list and probably into the larger tech noosphere too, as even in the context of all the AI madness, some people just keep building. That’s inspiring. Let’s join the party!
đź‘‹