I'm making a game. It won't be there most of the time.

I'm making a thing. It's not done, it might never be "done" in the way you'd expect a game to be done, and that's kind of the point.

I've talked about Mirror's Edge and my fascination with it before, and it pushes me to actually put something of my own on the table instead of just gushing about everyone else's.

It's not going to be online all the time

and no, that's not a bug I'm too lazy to fix

Most games want you logged in forever. Daily quests, live seasons, a little notification icon guilt-tripping you into opening the launcher. I don't want that, and honestly I don't think the story I want to tell works that way either.

So instead: it's event-based. The game shows up for scheduled windows, a bunch of people go through a chunk of story together, and then it goes quiet again until the next one. No treadmill, no pretending there's "content" when there isn't. When the people have gone through the story whole and the game is over, I will release the binaries to make it available for replay forever, provided you had the technical knowledge and patience to re-deploy a considerably-big project, maybe after a while the source code, this is not off the table.

No HUD. None. I mean it.

your health bar was never doing as much work as you thought

If you've read the Mirror's Edge post, you already know where I stand on this. No health bar bolted to a corner, no minimap turning the world into a chore list. If I have to break immersion to tell you something, that's a design failure on my end.

Everything you'd normally glance at in a HUD lives somewhere you actually go check on purpose. The world tells you what changed when it changes. That's it. That's the whole philosophy.

There's a name for how the menus work too

yes, I named it, no, I won't explain it yet

"No HUD" doesn't mean "no UI," it just means the UI has to actually earn its spot on screen. Whatever's left: menus, prompts, a notification, all follows one ruleset I've been quietly building out for a while now, under the name LECP: Low Effort-Cross Platform.

I'm not getting into what it is exactly or how it works in this post. Just know it's not an afterthought bolted on at the end. it is as load-bearing as everything else here, and it'll get its own post eventually.

Built for machines that remember UE3

not because I'm nostalgic, because I'm stubborn

I'm not chasing whatever's trending in graphics tech this year. If a game needs the newest card to look decent, I'd argue it doesn't actually look decent, it's just leaning on horsepower instead of having a style. Same lesson Mirror's Edge taught me: the how matters way more than the what.

So the target hardware is deliberately old. Not because I hate good graphics, but because working inside a tight box is usually what makes something look intentional instead of generic. Ask any artist who's ever had to work with three colors instead of thirty.

Constraint is important here.

Why now

This might end up being my graduation project, yes, I'm an AI Engineering student, so it's right to ask "How can this be your grad project?" Well, there will be utilization of ML technologies server-side to make the game really dynamic and interactive, however, no LLMs that control NPCs concept is considered at all, pure ML for pure fun in unpredictible ways I'm still polisihing.

This whole thing is either a great idea or the reason I don't sleep for a year. Possibly both. I've got a whole personal ruleset for how I want this thing to look, feel, and behave, and I'm not going to bore you with it here, that's a "figure it out when it ships" kind of detail, not a blog post kind of detail.

What I will say is this: it's the first time I'm building something with an actual philosophy behind every decision (maybe even overengineered at times), instead of my usual "winging it" workflow.

Whether that philosophy survives contact with reality is a different post, for a different day.