The shop that no longer exists

The post "Fourteen months" starts in June 2025, with me sitting down in the evenings to write a system monitor on a laptop from 2014. That is true, but it is not the beginning. Three or four months earlier I was doing something completely different, and I have never written about it here, because it did not work.

It was called HuckHub and it was meant to be an online shop.

A shop meant to carry very few things

HuckHub was not going to be another place with a thousand products. It was going to have very few, and that was the whole idea. Before anything reached the site it had to pass a filter. Check that it actually works. Check that it can be recommended to somebody with a clear conscience. Publish only after that.

A great many graphics were made for it, and even more plans. Plans were what there was most of.

Why I stopped

I worked one thing out about myself fairly quickly: marketing patter is not for me. Writing copy meant to talk somebody into something wore me out more than anything else in that project, and I was bad at it, because I did not believe myself.

What I actually wanted to do looked different. Build a tool that solves a specific problem for a specific person. Give it away. Test myself against something measurable. And talk about it honestly instead of advertising it.

The shop died on that thought, not on a lack of time. Thirteen projects before it had ended somewhere between seventy and eighty-five per cent, in that comfortable zone where something works on your own machine. This one ended because part-way through I understood I was building the wrong thing.

What it left behind

There is no public trace of the shop. Two things remained, and both matter more today than it ever did.

The first: I started writing in public. The earliest posts went on LinkedIn and were written to essentially nobody, because I had no product, no site and no readers. The writing began before the application, not after it.

The second is less obvious and I only noticed it a year later. The filter stayed. The idea of recommending nothing you have not verified went straight from the shop into the program: today it is a database of 340 hardware compatibility entries and a rule that no specification is ever guessed. The shop disappeared and the one good part of it did not.

A gap in the code that was not a gap in the work

For a period of up to three weeks I worked as a long-haul driver. I had a phone on me and nothing else. I did not sit down to code once.

It shows in the public changelog and I am not going to hide it: between version 1.6.8 on 17 February 2026 and 1.7.1 on 10 April 2026 there are fifty-two days with no release at all. That is the longest gap in the program's whole history.

In those days I planned, wrote and ran everything that can be run from a phone in a cab. I have no commits for it, because a commit happens at a keyboard. I am not claiming there was no break. I am claiming the break was in the code, not in the work, and that is a distinction a changelog cannot record.

What can be checked

Everything below was read from its source on the day this post went up.

Why write this at all

Because for a year I answered "how long have you been doing this" from memory, and memory counted sometimes from the first commit and sometimes from the first idea. Both answers were true in their way and neither could be checked.

So here it is: the code started in June 2025, and everything else three or four months before that, with a shop that did not work out. That is the full version and the one I am sticking to.