Skip to main content
SubZeroDev

Professional uncertainty since 2026.

Well… Why not?


The Projects Are Reproducing

· 5 min read

I have reached the point where I need a memory system to keep track of my projects.

One of those projects is the memory system.

I appreciate the elegance. I question the management.

Over the last few days, since I started working with REA and Ghidra, things have become a little ridiculous. Taking software apart turns out to be an excellent way to understand it. It is also an excellent way to acquire ideas I apparently cannot leave the fuck alone.

There is a dangerous distance between I wonder how that works and I could build my own version of that.

Mine appears to be about three seconds.

A perfectly reasonable inventory of the damage​

First, the file manager.

Or, to be precise, the specifications for a file manager. I am keeping that distinction because writing down how a thing should work does not magically cause an executable to materialize. Even Hell presumably has acceptance criteria.

Then a Windows desktop shell, with LiteStep as a reference.

Apparently, managing the files was insufficient. I also needed to consider replacing the bit of Windows surrounding them. A modest adjustment. Merely the desktop.

Then the media browser: something in the spirit of ACDSee 3.1, with images, video, and audio. That one is taking shape too.

There is something beautifully offensive about looking at an old application, remembering what was useful about it, and deciding that the appropriate response is another development project.

At this point, my relationship with the Windows desktop is starting to resemble a hostile takeover conducted entirely through specifications.

Meanwhile, in the games department​

I also have the online adventure-game work with Bulgarian-style graphics, growing out of the absurd adventures in the game engine.

Because obviously, in the middle of reconsidering the desktop, I needed somewhere to put the absurdity.

Most people put it in a group chat. I appear to require an engine.

And Hollow Moon now has a Windows desktop app that actually updates itself.

That last detail is worth pausing over. A working application. With an updater. Something I could potentially publish and put on Steam for five dollars or whatever price eventually makes sense.

Could. I have not transformed a possible Steam release into a completed commercial launch by mentioning it in a paragraph. The marketing department can sit the fuck down.

Still, this has moved rather a long way beyond a vague idea for a game. There is an application now. It runs. It updates. It has the sheer audacity to be a thing I might actually ship.

Naturally, I built something to remember all this​

MemoryKit is almost done.

I find that sentence increasingly entertaining.

I am building persistent memory while generating fresh work quickly enough to make persistent memory feel like emergency infrastructure.

The system is approaching completion. Its first stress test appears to be its owner.

Somewhere in this mess is a sensible engineering instinct: preserve context, keep track of decisions, stop rediscovering what happened. I stand by that instinct. I would simply like to know why exercising it also results in another repository.

Every individual decision is defensible. That is the really irritating part.

A file manager? Useful.

A shell? Interesting.

A media browser? Absolutely.

Games? Entertaining, and potentially something other people could play.

A memory system? At this point, practically self-defence.

Put them together and the defence begins to sound like evidence for the prosecution.

I asked for an inventory. That was a mistake.​

Eventually I asked the obvious question: what am I actually working on, and what have I forgotten?

Seeing it assembled was substantially more absurd than carrying it around in separate mental compartments.

In my head, each project occupies its own perfectly respectable little room. Put the rooms on the same floor plan and I appear to be constructing a fucking palace.

Some rooms are drawings. Some have walls. One has a working updater. I am not claiming the palace is finished.

I am saying the architect has lost track of the wings.

The recurring problem is that solving something exposes another useful thing to build. Then that thing wants a reusable foundation. Then the foundation becomes interesting enough to deserve a name.

By the time I return to the original problem, it has acquired neighbours.

This is scope creep with a breeding programme.

And yes, I find it hilarious. I am losing track because there is actual movement: specifications becoming clearer, applications taking shape, games becoming playable things, infrastructure getting closer to useful. The stages differ. The activity is real.

That is why I want this written down. A journal seems an appropriate place to record the moment I noticed the projects had started reproducing.

For now, I have a list.

I am resisting the temptation to give it a plugin architecture.