Welcome in the year 2012. Zini already wrote the summary of openmw status after 2011.
Well, 0.12.0 is quite close now. Jhooks1 is probably amazing himself (not me, I’m already confident about him) with his developing skills. To quote zini:
I am away for just a couple of days and suddenly tons of stuff happens.
Improvement is really significant and a large step in the right direction. It seems that some other problems left are related to drivers and libbullet so suffering level (this sounds so cool) depends on the version of driver/lib.
And there is almost nothing we can do with it besides reporting bugs. K1ll reports that actually it WORKS since the bullet bug seems to be a fixed in the last svn revision. And that nasty bug could cause quite massive fps drops.
GUS works on journal. It works but still needs some love, quoting GUS:
Text is not properly aligned to the book and the book is not properly aligned/scaled on the screen.
… and we have *.deb packages! Thank you BrotherBrick! At the moment I can’t see them under downloads section but this is probably a matter of time. This packages where build for 32 and 64 bits ubuntu.
And jhooks found the major problem behind poor performance of animations? Animations runs, faster reaching the openmw standard (read as “runs still slow but not so terrible slow as before”)
Zini finished both 0.12.0 tasks that was reserved for him. Doing 200% while other developers slips away makes him almost heroic. Well, now zini is doing code maintaining stuff in the engine class.
Bullet performance. Although k1ll problems where caused by bug in the bullet itself GUS found idea to increase physics performance? I’m not sure if we will get better FPS since this is probably not a bottleneck now. At least It may help in future to speed rendering up.
jhooks was able to make hands open and close. And it’s the good news. The bad news is that with all this animation stuff performance is quite poor.
Solving performance problems is not strait forward but in theory It looks quite simple: find the bottleneck, part of procedure that runs slowest; then try to fix it. If needed: repeat. GUS already demonstrated this tactic. The worst thing that we can discover is the bottleneck function that can’t run faster but investigation just started and nothing like that must happen.
But at least It works! Keep this in the mind all the time.
Bullet optimisation also slowly moves forward. K1ll still suffers poor performance on his system and helps GUS to find the problem. The situation is not clear at the moment, this can be bug in the bullet itself.
OpenMW 0.11.1 is late because We lack linux builder. There are no word to express what I think about it.