The requirements for the v1.0-pre3 release of OctoberPlains have been met so I posted the release to GitHub.

The most noteworthy change in this release is the introduction of a sky light mechanic and a day-night cycle. Some small tweaks to the worldgen were made, as well.

I also updated the roadmap to include a pre-release for some placeholder audio. This is mostly just to make sure that I have a sense of how this works within LibGDX.

This means that work on v1.0-pre4 will now begin. This release is all about network optimization, stress testing, versioning, and hardening. It should be interesting to design the kind of test rig I will need for this: Attach hundreds of clients to a server and have them do... something.

As for v1.0-pre3, the game is now playable and not terrible. The main problem is the constraint defined for the OctoberPlains project: The perspective. It is pretty hard to see what you are doing when walking down a tunnel since the walls are dark (the cursor over them shows what they are, though). I was thinking about fixing this but it really isn't the point since this release isn't about being a great game, but a great test-bed for the core technology which will become the basis of OctoberPeaks, in a few months.

Also, as the content-centric demos and pre-releases are now behind me (graphics and audio aside), I am starting to think about how the transition to OctoberPeaks will actually work and if any of that support needs to be pushed back into this timeline. Mostly, this is in regards to movement as you currently can't walk diagonally, which will absolutely need to change for the OctoberPeaks interface. I suspect that this can probably be changed before the release-candidate, as I want to minimize the number of commits from the last OctoberPlains baseline to the first OctoberPeaks baseline.

In a perfect world (which might be possible), both clients could coexist on the same server during a period of logical overlap. While a bunch of things will need to change in response to the different interface (more ray-casting and the introduction of "facing" for an entity), I suspect that the vast majority of it should start out compatible.

As for now, I am looking forward to pre4 and pre5 since they are mostly down in the guts of the system, measuring core behaviours and performance, coming up with theoretical maximum capacity estimates, etc. While sketching out and prototyping the initial framework of how a large system needs to be designed is probably my favourite part of this profession, being able to do serious analysis of something you have built is probably my runner-up.

So far, so good,
Jeff.