Someone mentioned to me on a call recently that I hadn't updated taro.codes in over a year. It hurt a bit, but honestly, the thoughts page does make it look that way. Behind the scenes, though, I've been changing quite a lot during 2025 and early 2026.
A few of the more visible improvements:
- Redesigned and rebuilt the creations and books pages from scratch. I put a ton of effort into this, and I must say I'm really happy with the result.
- Added a vernacular section, so I can keep track of expressions like "software archeology" and words like "frobnicate"
- Client-side navigation1, which improved the UX of the whole site!
- Added a "user preference" option to the theme toggler (
prefers-color-schemeis respected) and keyboard navigation
Under the hood, I migrated from Netlify to Cloudflare Pages2, finally got rid of Sass in favour of native CSS, added Shadow DOM where it made sense, and reduced the dependence on Eleventy and Nunjucks/Liquid.
I started a few articles too, but life happened and I never got to finish them. 2025 was a rough year.
Part of why taro.codes was quieter over the past year is that I've also been building other things. I spent a lot of time on Jolteon —my biggest project to date3— and on Vaporeon (currently private).
I'm still figuring this out, but taro.codes is much more than a blog or collection of essays. For now, I think I'll post updates like this, or even quick, small thoughts, and not just long-form essays.
- After the Navigation API got to baseline, I decided to give it a go and implemented client-side navigation (a very primitive version of it). Kinda like what Astro does under the hood, but manually implemented in ~100 lines of code because... well, because I can (: here's the source, if you're feeling curious: navClientSideRouting.js, secondaryNav.js.
- And completely broke the newsletter subscription flow in the process, right after I got my first subscribers. Obviously, it's still broken.
- Over 800 commits and almost 9k lines of code, and even 2 fixes to upstream projects!