Issue #3: My Way

🎶 My friend, I’ll say it clear, I’ll state my case, of which I’m certain. 🎶

This time, it’s all about opinions. There are plenty of them in the world of metaframeworks – both good and bad. The important thing is to choose wisely amid the noise of biased voices and cargo cult gospels. Today, we’ll dig into some of these opinions from all possible sides and standpoints, as usual.

The Good

First, I’ve got a couple of 2024 retrospectives for you from two of the most dynamic and clever metaframeworks in the ecosystem.

The Astro team (notably, this time it’s written by a community member) came up with their “Year in Review” piece. I’ve already mentioned in one of the previous issues that Astro had an incredible year, and they have a lot to be proud of. They do a lot of things “their way,” and the funny thing is that their approach often gets adopted by other frameworks. They’ve earned their place as one of the most loved and innovative technologies in the JavaScript world last year.

Zach Leatherman summarized 2024 from Eleventy’s perspective. It was a huge year for the team – a new major version, an acquisition, and a conference. While Eleventy might not be the loudest framework out there, it’s one of the most thoughtful and consistent, which is crucial in this dynamic metaframework ecosystem. Not to mention their new awesome certification for developers who’ve truly mastered Eleventy. Yours truly was one of the first Certified Eleventy Super Professional Web Developers 2025, for what it’s worth.

And speaking of yours truly and the metaframework ecosystem, I’ve published the first draft of my big project: Metaframeworks Comparison. It’s an attempt to summarize and analyze the current state of the ecosystem, including pros, cons, features, and more. Check it out and let me know what you think – you can simply reply to this letter with your thoughts and ideas.

The Bad

Now, for some bad news. Metaframeworks can get toxic. As with everything in this world, the truth lies in the eye of the beholder, but here we’ll take a look at some thoughts on why metaframeworks can be suboptimal in certain software development scenarios.

Seasoned web developer Mánu Fosela shares his opinion on the issues with metaframeworks and the actual business incentives behind them in his piece titled “Over-Architecting in Web Design”. This isn’t a conspiracy theory but rather an attempt to find a sane balance between technological necessity and mental overhead.

The team at ComfyDeploy brought a case study with the explosive title “You Don’t Need Next.js”. It’s a somewhat edge-case but thought-provoking exploration of what metaframeworks (when misused) do to us and our developer experience. Skill issue? Decide for yourself, but there are a lot of interesting findings leading to opinions like:

Next.js is great for what it’s meant to do – landing pages, SEO stuff, that kind of thing. It makes simple concepts look fancy (Server Actions are just REST API calls).

The HTML All The Things podcast recently explored a big question: “Why Is Everyone Unhappy With JavaScript?”. To be honest, that’s an obviously false assumption, but as it’s a survey analysis episode, false positives and negatives are expected. What’s true is that JavaScript fatigue is real, and metaframeworks play a significant role in it. Why? Check it out – it might be insightful.

The Noteworthy

LogRocket’s team of writers regularly produces great content on JavaScript and web technologies. This time, they address the topic of metaframeworks, specifically comparing three of them – Analog, Next.js, and SolidStart. This isn’t just a surface-level theoretical overview but goes a bit deeper: some practical aspects are reviewed and tested to help readers form their own opinions based on their experience, goals, and preferences. Analog isn’t the most frequent guest in such overviews, so I’m personally very happy to see it included.

Addy Osmani, one of the greatest wizards in the web dev community (who’s been diving into AI stuff lately), came up with a detailed comparison of v0, Bolt, and Lovable. These are some of the most popular LLM tools for rapid full-stack prototyping in your browser. I’m particularly interested in this because Bolt and Lovable are getting better at generating cool metaframework web apps every day. They still have their flaws, and the experience can be daunting at times, but with advice like that in this article, your experiments can definitely improve.

And speaking of hints and tricks: the experience of playing with AI prototyping tools is another reminder that official documentation can be incredibly helpful. For instance, Lovable has a great page with tips that can be useful not only if you’re using Lovable’s web interface but also with other similar AI-based tools. The same goes for Aider – a popular and robust Python-based tool with flexible LLM support – their docs also have a lot of insights.

All in all, don’t always trust opinions—think for yourself using the best information you can find. Hopefully, some of the resources in this issue will help you with that.

👋