Day 16 of my solo journey to make my engineless mmo.

Very boring day, yesterday I ended my day doing some benchmarks on a few ideas I have for application-layer DDoS mitigation, I'm quite happy with a couple of them.

Today I spent half of the day working on a small static site generator that works on a custom format I've used for a few days to track stuff I want to do and stuff I have done in plain text. The other half was spent mostly working on stuff like setting up client and server loops, building the clock synchronization system, building some data structures to store stats, implementing a tiny platform layer to have monotonic nanosecond clocks and futex in windows/linux/macos...

There are still a few pieces that I have to make before having it playable online, I always like to deploy my servers with hot-reloading, but I think I'll add that to the backlog because I want to move back into more visual tasks somewhere next week, since I feel very sad to just dump something here without a couple images.

#woao #solodev #gamedev #indiedev #mmo #devlog #enginedev

Betawatch: Star Wrath enters closed beta, Time Takers gets signups - and more #MMO testing news from the week behind us!
🏄‍ https://massivelyop.com/2026/02/20/betawatch-star-wrath-enters-closed-beta-time-takers-gets-signups

Day 15 of publicly logging my solo #gamedev journey to make my #mmo.

Yesterday and most of today I didn't really work on the game that much.
More specifically I have been hesitating and researching about how to tackle the main (gameplay) protocol of the game, the login, and the main spatial data structures.

This may be controversial, but while I'll use proper crypto (X25519 kex) for the session setup, I will go with (mostly) unencrypted gameplay datagrams. Please don't yell at me, I am not going to claim I am not crazy because I'm working on an MMO. But I truly believe that authenticating using a keyed blake2b on top of regular UDP should be good enough for the game I want to make.

Also, since I'll be going the buffer history with delta compression approach, the sequence id and cumulative ack both will allow me to use smaller nonces while still preventing replay attacks.

Day 14 of publicly logging my solo #gamedev journey to make my #mmo.

Yesterday I continued with the project management stuff in order to be able to track a bit better how I spend time, so far amost half of the time would have been saved by using an off-the-shelf engine, let's see how this evolves with the weeks and months.

I went on to add some quick shading and batching to the raylib prototype just to keep it playable with a large number of colliders. It turns out the gains were bigger than expected since each frame no longer needs as many CCD iterations. This is nice and something good I did not account for when optimizing the previous day (I mean, just look at that no-broadphase single-core physics time in the top-left corner, for that many triangles, this is why I'll never understand people complaining about a few branches here and there in a game like Yandere Simulator).

More importantly, I also implemented a pretty decent "first version" of the character controller using platforming games techniques (all of them boild down to buffering of state/input and using non-realistic, contextual physics).

For now I have left those techniques somewhat exaggerated since it will help me test and validate stuff, but it already feels much greater to move around than before. Unfortunately the most immediate improvements that could be made rely on rendering and graphics, so that will have to wait until after I get it to work fine online, I am pretty sure I could get that done today, but I never considered how much harder it is to keep working alone.

Bug-wise I already have a bug#2, but I won't be fixing it immediately like I did previously. More specifically spheres will try to slide in geometry edges when jumping into a corner. It's not a deal-breaker and not a blocker, so I'm sure I'll think of a cheap (compute-wise) solution to this eventually.