I'm sorry. I started work on rewriting the netcode from scratch and it looks promising, but it's gone stagnant for several months now and may never be finished.
This
was JTE's ridiculous roadmap of things JTE wants to do to "make SRB2 great again":
- Netcode rewrite so that one player lagging doesn't lag the entire server and gameplay itself is framerate-independent. No more control lag. No more resynchronization pauses. This will probably break the majority of how Lua scripting works currently, since it represents a very major paradigm shift of how the game works on a core level.
- OpenGL rewrite to use OpenGL 2.0 or higher instead of the ancient fixed pipeline. Primarily so that we can use GLSL shaders to make OpenGL match what Software mode can do much much better, even if it isn't flawless. It should be capable of becoming the primary rendering mode. (I was actually relieved when Furyhunter began work on OpenGL, even though he didn't believe me when I said it'd need a complete rewrite to work out, but then nothing became of it. We gained nothing.)
- All of the game logic rewritten to support a higher (or even uncapped) framerate natively. After the netcode is framerate-independent and desynchronization-proof and a fast and efficient hardware renderer is suitable to be made the default, there would no longer be any reason to have Sonic thokking along at an absurd 60 pixels-per-frame at a weird and choppy low framerate like 35, when he could be moving buttery-smooth 35 pixels per frame at 60 frames per second, the same momentum but with twice as easy movement for your eyes to track and respond to.
Naturally, all this equates to essentially rewriting the entirety of what little is left of the original engine from the ground up, and also hacking around our 16+ years of "35 frames"-based logic, which permeates absolutely every aspect of the game, from the animations to the physics itself. It is absurd, cannot possibly happen all at once (expect one of those things every-other major version,
in the best-case scenario), and may not happen at all (ridiculous bugs can crop up, massive amounts of things need to be rewritten, and there's few programmers talented enough to stomach the task)
It... It's not worth it. I am not in the sort of position in my life where I can dedicate my time to a project that's this large of a scale with only minimal returns.
Mind you, these aren't the only things that would happen in SRB2's future. Better mapping tools and more content are being added all the time, and Lua scripting is continually digging deeper into more parts of the core engine to make literally anything you could want to mod in become possible (even if it has to be a little hackish...)
Given that, I believe most shrinkage of SRB2's community can be attributed to:
- Nothing "exciting and new" to report in a long time (You can only say "SRB2 released a new version and is still awesome" so many times before it gets boring. Even "SRB2 finally gets slopes!" is rather bittersweet when it's so many years late.)
- Incompatibility with modern operating systems (Windows 10 required us to move the software rendering engine to be piped through SDL2 and turned into an inefficient hardware rendering engine by default, just because it can't handle ye olde DirectDraw at all anymore)
- Or because Sonic the Hedgehog itself is falling in popularity, Sega long having moved on to simply licensing the character out to make an easy buck rather than actively striving to produce quality videogame content themselves.
Either way...