Poll: Kickstarter for art revamp for AI War? (Tip: Vote no)

[LintMan]

Some thoughts:

- HUD improvements might be nice, but even there, I'm desiring more in terms of UI improvements than actual graphics changes.

- Nicer planets/nebulas might be nice, but I barely pay attention to that stuff once I'm mid-game.

- Nicer ships would be cool, but they look fine now, and I'd really worry than more detailed ships would increase the memory/processing power requirements and slog things down in big epic games.

- which brings me to this idea: I'd gladly kick in for a Kickstarter to do a 64-bit version of AI War or at least something to free up some of the current memory restrictions if it isn't the 32-bit 4GB limit.

[TechSY730]

- which brings me to this idea: I'd gladly kick in for a Kickstarter to do a 64-bit version of AI War or at least something to free up some of the current memory restrictions if it isn't the 32-bit 4GB limit.

The only problem is that the memory issues are NOT due to the 32-bit address space restrictions, but rather due to the rather questionable garbage collector implementation of the engine that causes it to crash if it has to clean up too much. (Around 900MB heap usage is about where you really start to risk tripping the crash)
Sadly, this is outside of Arcen's control, it is a bug in the version of Mono that the engine, Unity3D, uses for this version.

I would rather the money go to Arcen to give to Unity to motivate them to fix up this bug (by migrating to a newer version of Mono; later versions have far, FAR better garbage collectors) and their Unity 4.0 bugs that cause very, VERY bad performance on some systems, which is possibly the only thing holding Arcen back from moving to Unity 4.0 for AI War. (Unity 4.0 would also implicitly come with native Linux support )
And then, as a stretch goal, the money go into at least trying to make the engine more parallelized, which may require tweaking the network model some. All-in-all, not an easy task at all, which is why it would be a stretch goal.

[Aklyon]

Isn't the unity 4 thing mostly a case of 'we agree with you it could be much better but we need a simpler test case than an entire huge game to test with'?

[TechSY730]

Isn't the unity 4 thing mostly a case of 'we agree with you it could be much better but we need a simpler test case than an entire huge game to test with'?

Yes, but they have been having trouble making such a "simpler reproducing case", and they don't really have the money to afford spending more man-hours on this than basically an "on the side" level. (Or at least, this is what I recall being told)


It would be nice if someone on the forums more familiar with Unity could take a stab at making a reproducing case. (I know C#, but not Unity)

[x4000]

I think this puts you in an unsustainable situation in the future. Right now Keith is very agile in implementing gameplay changes and making the game more fun. This isn't a case of ASCII art interfaces or anything.


Let this thread die.

Agreed on all points.