My communication has been bad. I'm gonna try a new format. Two types of posts: random progress updates whenever something's ready, and monthly wrap-ups. My brain is happier with less rigid structure, but consistency is important, so this might strike the right balance.

What's the state of things?
This year was low in visible progress and lacking posts. It's because of a major life upheaval. I thought the feelings of chaos would eventually  calm down and let me work, but instead I've just become accustomed to operating within it. As 2024 ends, I'm left with feelings of satisfaction and embarrassment.

I'm very excited and satisfied with the state of the engine. It's a little daunting in size when looking at the whole thing: a lewd-focused sandbox game engine needs a lot of custom, hand-built things. But despite its growing size, I think each system (fluid dynamics, sensations, emotions, reactions, overlapping animation management, modular body parts, etc etc there are a bunch) is in a workable enough state to be impressive on its own, if it were the focal point of a game. Not only that, but my programming skill is so much better after all the effort to remove stress from my life. I'm even better at making this stuff than ever.

But here's where I'm embarrassed. I'll be blunt. I'm heavily neuro-spicy. I have had a tendency to lose perspective in ways that might seem silly to others. I've got this ridiculously powerful engine already. But I've left out a couple relatively simple pieces that it needs to actually... run a game. Thus I'm stuck just methodically trying to fill those missing holes in the engine while my audience waits with straining patience. That wouldn't have happened if I'd just built things rougher and more big-picture instead of being all meticulous about doing it piece by piece.

My work has gotten less rigidly focused over the course of this year. I'm just hopping between interconnected things. Right now it's a much-needed UI code upgrade that's helping me test and finish off the reactions code, and will also finally give me a simple UI to do everything else with.

In hindsight I would've opted for simpler systems for everything at the start, enough to run something very basic. Then released a whole little game with it. Then built the engine more. Repeat.

I regret that I didn't do it that way. Back when I started remaking this engine, I was so frustrated with how the old engine felt like it was at its limit before I even got half of my ideas into it. And I knew the reason it was so limited was because of my past mistakes. My early choices as a less-experienced programmer had led to a tangled and unstable foundation. So I tried to write this new engine the way I wished the old one had been made. I was too afraid of being limited by my new engine, and that led me to focus on too much depth first. 

In the long run, it's fine. I'll finish up the core pieces needed, and I'll show stuff off as it's ready. The depth I put in isn't wasted, but it's very inconveniently timed. Sometime in the relatively near future, I'll have the missing pieces filled in and ready to play with.

Those of my supporters who decide to keep funding this process, thank you. Those who don't want to, thanks so much for your support in the past. Genuinely. Even if this project ended up losing most of its funding, I'd finish making it, albeit much slower since I'd need other work. It's just too cool. I gotta see it through.

My next post will contain more specifics about what's being built n stuff. Thanks for hanging around throughout all of this <3

Whoops, last two posts weren't set to public. Substar makes it very easy to accidentally make posts paid-only. All posts here should be public except subscriber-only releases.

Quick update since I'm only partway through this.

To showcase how powerful and modular this thing is, I've given each eye a different reaction to the same emotional data in the following example, with 3 layers of reactions balancing with each other at once.

Observe:

It's balancing between 3 expressions- the base layer (neutral), and primary expressions to represent the 2 emotions I'm playing with. It can balance as many expressions as needed for each individual piece of the face and body. Also, when there are lots of layered emotions, it'll combine the weakest emotions into one simpler basic emotion to display. That's the hard part, done! 

I'm still debugging the latest addition: quick-swapping groups of assets using the reactions system. You can see it in action a little here with the mouth and eyebrows, but there's some more to do. 

Plus, I'm partway through a GUI to edit expressions, which'll speed things up a lot~

That's all for now! This month was weird, but the progress is actually looking pretty great. Each step in the game's todo list becomes easier than the last, thanks to all the engine work. Pretty soon I'll be able to add in a ton of expressions and body animations without touching any code. Looking forward to that ^^

Major breakthrough! Simple and complex reactions finally work! I'm inconveniently sick right now though, so gimme a minute to put together a visual presentation of it.

The system functions exactly how I thought it would, minus a few bugs. It's a modular chain of brain, body, and animation logic that can be as complex and nuanced as you need it. So, for things like blushing, it's very simplistic. For expressions, it's a monster of brain inputs and face configurations.

But... I was using stress-testing animations while making it, so it kind looks like a FNaF horror creature right now. When I'm less sick, I can swap in some real animations and show you why this system was worth the effort~

<3

edit: this post was accidentally set to the paid tier instead of free. My bad >.<'

So last time, we had an unexplained fps drop. I showed off the two separate debuggers I made for the logic and visuals in the game. The problem causing the fps drop turned out to be tricky though, so I developed both of the debuggers more. Now they both give me a nice simple overview of what's happening in the game.

The animation renderer was slowing down. I was telling the debugger code to look for a big problem, but there were no errors, and the animation logic wasn't actually taking up any extra processing. And yet the fps would drop.

Since the code couldn't catch the problem, I added more human-readable stats to my debugger. The key was getting the debugger to tell me- how many times are we asking the animation code/renderer to do specific types of tasks? Once I implemented that, suddenly everything was way easier.

This is a bit hard to explain, so here's a couple of screenshots showing how I was trying to debug this before (Method 1), and how the debugging works now (Method 2).

That's the animation logic asking the animation code to start a new animation 30k times. Meaning there are 30k animations all layered over each other playing all at once. Whoops. You'll notice that the animation logic isn't actually causing any lag (the average/total durations are super small). So it didn't show up in performance checks.

Yep. Now I have a simple readout of everything happening in the game. If something is causing performance problems, but the debug code isn't able to tell me why, I can now just look at the readout and see if anything's weird. 

Turns out I only needed to add 2 lines of code to stop the Body Reactions system from infinitely starting new animations. So yay, that's fixed! This debugging solution looks like it'll be good enough to carry us through at least the first few releases, if not much further.

Moving on! The Brains are throwing some errors, so I'm off to fix that next.

While checking brain reactions, I noticed a big uh-oh.

A gradual linear fps drop. I figured it should be no biggie. It being gradual indicates something that should only happen once is instead happening over and over again, slowly eating up all the processing power. 

Then I realized I had no idea where in the process this was happening and no way to check except... manually, system by system. Huuuge waste of dev time.

So I built some performance profiling into the engine's debugger. I made a profiler for both the game's visuals and logic since they're two different systems. 

Neat! Now I can see where any unexpected processing or lag is happening.

A side effect is that I can see when something is processing unexpectedly, but isn't causing lag. Turns out a bunch of game systems were sneakily running when they didn't need to be.  Thanks to the new Entity Component System structure of the engine, I can easily look and see why each system may be active. 

The effort to make these profilers has already paid off twice as much as I expected. Should've had this from the beginning. Now that I can see what's happening, it's back to debugging reactions for me.

That's all for now! 

<3