Thinking Out Loud

Getting the Languages Back

Anthony Shead/September 3, 2026/5 min read
Getting the Languages Back
fig. 1 - Getting the Languages Back

Four projects, four languages

There are four new folders sitting next to the portfolio, and the interesting thing about them isn't what they do. It's what they're written in.

Translator Python. You type something in, you pick a language, and it gives it back to you in that language. A local Python install, nothing hosted. The simplest of the four and the one closest to finished.

Alarm Clock C++. C++ was always the one I trusted for anything numeric, and I hadn't touched it in a while. Starting with something familiar felt right one project where the language isn't the obstacle.

To-Do List C. I have written very little C, and the little I did was a couple of years ago. A to-do list in C is a deliberately silly choice for something that's ass of now seems easy. That's the point. Manual memory management for a problem I already understand means the problem isn't the thing I'm learning.

Calendar Rust. New language, and the one that'll take the longest. Calendars are also quietly nasty recurrence rules, and time zones eat more time than anyone budgets for. New language plus a genuinely hard domain is probably the wrong order to do things in, and I'm doing it anyway.

So: two languages I'm recovering and two I'm effectively starting from scratch in. That split wasn't accidental. I used to code in C, C++, and some Python, and I want to know whether that's still true or whether it's just something I say. The familiar two are the check. The unfamiliar two are the actual work.

The honest risk with running four at once is that you end up mediocre at four things instead of good at one. I've thought about that. Right now I'm in a stretch where a lot of my time goes to looking for work and finding clients for SorenLab, and that's not time that keeps coding skills sharp; it's a different muscle entirely. Four small projects in four languages is the thing I've got that pushes back against the drift.

The main site

The main site is live. Nav across the top: My Work, Case Studies, About, Blog, Connect, dark mode toggle on the right. Below the fold, there's a globe.

The globe is the part I care about, and it works. You go to Connect, you type your name where it asks if you want to be put on the globe, and it lands in a queue. I get it in the admin view as a new submission, and I either accept or deny it. On accept, the name goes onto the globe and stays there. The pins on it now are real people who did exactly that.

The one thing worth saying outright: the placement is random. Names get dropped at a random latitude and longitude, not where the person actually is. There's no geocoding behind it, and I'm not collecting location. It's a guestbook that happens to be shaped like a planet.

I like this for a reason that isn't really about globes. A portfolio site has to do two jobs that pull against each other: it has to be proof of work, and it has to be a piece of work. Every hour spent making the site itself impressive is an hour not spent on the thing the site is supposed to be pointing at. The globe resolves some of that tension because it isn't decoration. It's the one thing on the page that actually does something, and it's small enough to finish.

The four projects are still under construction. The site is the finished thing pointing at the unfinished things, which is a slightly uncomfortable arrangement, but it's an accurate one.

There's a third thing in the Figma file that hasn't started yet: an architecture portfolio. Sketches, builds, the design work that isn't software. It's mostly an empty canvas right now. I'm mentioning it because it exists and because I'd rather say what I'm planning than surprise people with it later.

Figma

I've been learning components properly for the first time.

In practice, that meant building a small library of reusable shapes instead of redrawing things: folder icons in two states, a plain document, a couple of notebook variants, a shelf element. Boring work. The kind of thing you skip when you're in a hurry and then pay for later, when every icon on the page is subtly different from every other icon and you can't say why the layout feels off.

It didn't take as long as I expected. Design isn't new to me; I came up through Adobe, Photoshop, and Illustrator, so the concepts transferred. What was new was the discipline of building the piece once and then using it, rather than building something that looks like the last thing I built.

Here's the part some people have already seen without knowing it. Those components are what the Creator Store images are made of. The Vault product cards Starter, Intermediate, Advanced, Master, Organizer, Builder, Integrator, Specialist, Expert, Architect, Grandmaster, Sovereign are all assembled from that same set of shapes. Different arrangements, different color treatments, same underlying pieces.

That was the payoff I wasn't expecting. I started learning components for the portfolio site and came out of it with the asset pipeline for an entirely different product. Twelve product images that would otherwise have been twelve separate drawing sessions came out of one afternoon of setup.

This is the second time in a month that the boring version of a task turned out to be the one that paid. I don't have a clean lesson to draw from that yet, but I've noticed it.

What's next

The next focus is a more serious UI for the four projects. Right now they're functional-at-best, and they look it, which is fine while the language is the hard part, but won't stay fine. There's a reason I want them to look like real software rather than exercises, and I'll get into that properly soon.

New posts, in your inbox

Roughly one a week. Confirm once, unsubscribe anytime.

Writing like this is reader-supported.Support on Patreon →

0 comments

Read next