I love a good exposé on how a front-end team operates. Like what technology they use, why, and how, particularly when there are pain points and journeys through them.

Jim Simon of Reddit wrote one a bit ago about their teams build process. They were using something Rollup based and getting 2-minute build times and spent quite a bit of time and effort switching to Vite and now are getting sub-1-second build times. I don’t know if “wow Vite is fast” is the right read here though, as they lost type checking entirely. Vite means esbuild for TypeScript which just strips types, meaning no build process (locally, in CI, or otherwise) will catch errors. That seems like a massive deal to me as it opens the door to all contributions having TypeScript errors. I admit I’m fascinated by the approach though, it’s kinda like treating TypeScript as a local-only linter. Sure, VS Code complains and gives you red squiggles, but nothing else will, so use that information as you will. Very mixed feelings.

Vite always seems to be front and center in conversations about the JavaScript ecosystem these days. The tooling section of this year’s JavaScript Rising Stars:

Vite has been the big winner again this year, renewing for the second time its State of JS awards as the most adopted and loved technology. It’s rare to have both high usage and retention, let alone maintain it. We are eagerly waiting to see how the new void(0) company will impact the Vite ecosystem next year!

(Interesting how it’s actually Biome that gained the most stars this year and has large goals about being the toolchain for the web, like Vite)

Vite actually has the bucks now to make a real run at it. It’s always nail biting and fascinating to see money being thrown around at front-end open source, as a strong business model around all that is hard to find.

Maybe there is an enterprise story to capture? Somehow I can see that more easily. I would guess that’s where the new venture vlt is seeing potential. npm, now being owned by Microsoft, certainly had a story there that investors probably liked to see, so maybe vlt can do it again but better. It’s the “you’ve got their data” thing that adds up to me. Not that I love it, I just get it. Vite might have your stack, but we write checks to infrastructure companies.

That tinge of worry extends to Bun and Deno too. I think they can survive decently on momentum of developers being excited about the speed and features. I wouldn’t say I’ve got a full grasp on it, but I’ve seen some developers be pretty disillusioned or at least trepidatious with Deno and their package registry JSR. But Deno has products! They have enterprise consulting and various hosting. Data and product, I think that is all very smart. Mabe void(0) can find a product play in there. This all reminds me of XState / Stately which took a bit of funding, does open source, and productizes some of what they do. Their new Store library is getting lots of attention which is good for the gander.

To be clear, I’m rooting for all of these companies. They are small and only lightly funded companies, just like CodePen, trying to make tools to make web development better. 💜