OpenMW 0.33.0

[Zini]

A little late, but here it is anyway: The 0.33.0 posting.

It seems we reached our goal from last time. We are mostly feature complete. There are still a few bits and pieces missing, but nothing essential. Of course there are still plenty of bugs left, but that is a different topic.
Everything going well on the OpenMW side, I would say. Nothing special that requires highlighting.

Surprisingly the development of OpenCS is finally also picking up speed. We had a couple of new contributors (including contributors moving their attention form OpenMW to OpenCS). Most of the important tasks currently on the tracker have already been picked up or at least people have announced their interest in them. Therefore I don't have anything special to highlight for OpenCS either at this time.

We already have over 50 tasks down for 0.33.0. Looks like another big release. Over the last couple of releases a release-cycle of about two months has served us well, so we should stick with that for now.

[raevol]

Anything we are working on for optimization at this time? Most of the feedback from our last release was about FPS, which is awesome because people are actually getting into the game and trying to play, but is a bit of a bummer, because they are being turned away by poor performance and not continuing to play (read: bugtest).

[Zini]

Nothing on that from my end. There were some minor improvements and it is possible that someone will come up with some more of that, but I don't expect any big changes until we are tackling the OGRE question (one way or the other).

Edit: I am only speaking about the default esms here. There is still the batching issue (we aren't doing it that well). For the default esms that shouldn't make a large difference, because most objects in these are statics anyway, which are batched okay-ish. For other a bit more interactive content (*cough* Redemption *cough*) I expect a significant performance increase.

[scrawl]

Shouldn't we aim for faster release schedules, now that (by measurement of bug reports) many people are already legitimately playing OpenMW?

[psi29a]

Just a heads up, I've pushed a patch to Debian (which will shortly be live in Ubuntu) for Ogre3D that enables the GLES2 renderer option. This is in relation with my work on the RPi.

This means that on Debian and Ubuntu, we'll now have two options to choose from in the launcher. The GLES2 renderer is smaller and more efficient than the standard OpenGL renderer in Ogre3D.

@Scrawl: Do we currently take advantage of anything above OpenGL 2.1 (GLES 2.0) ?

[scrawl]

Just a heads up, I've pushed a patch to Debian (which will shortly be live in Ubuntu) for Ogre3D that enables the GLES2 renderer option. This is in relation with my work on the RPi.

This means that on Debian and Ubuntu, we'll now have two options to choose from in the launcher.

Interesting idea, but that is a problem. The GLES2 renderer hasn't been tested at all (except on Android, where several bugs cropped up that may or may not be Android specific). I don't think we want to introduce a visible choice like that when we can not ensure that it will work properly. Also, it will not work (read: crash) for drivers that don't implement EGL (on desktops, I think only Mesa drivers offer EGL right now).
That's why we should disable this choice in the launcher (but of course, the choice is still there by modifying the settings.cfg file manually).

[Zini]

@scrawl: I don't see anything critical in the remaining releases up to 1.0 (maybe except for optimisation and of course OpenCS). Therefore I also don't see a reason for faster releases. The one month cycle we had before was kinda nice, but it was also pretty rushed. I would prefer if we take our time.

[Tinker]

Taking time and getting it right is much more preferable to rushing and letting things slip through. Of course the the preference is to rush it and get things right

[scrawl]

scrawl: I don't see anything critical in the remaining releases up to 1.0

I don't agree with this. There are still many gameplay critical issues remaining. For example 1856, which makes creatures that were supposed to use spells way too easy to defeat. This is already implemented in master and I would prefer not to make it sit there for 2 months.
AI is also still a big problem (see the recent thread about it) and fixes in that area should be highly welcomed by players. 1472 can place the player out of interior bounds and will need console 'tcl' use or loading an earlier save to fix it.
#1125 (AI fast-forward) maybe not one of the more prominent features, but its absence still makes some thief builds useless (not being able to "wait" for guards to go elsewhere)
#1875 - Actors in inactive cells don't heal from resting. Fixing this is very important because again, the game is made way too easy.
#1113 - NPC door AI. Currently you can close the door behind your enemies and leave them there forever.
#1374:

Issue: The AI in OpenMW will keep attacking you underwater until they drown themselves. This means that you can aggravate an entire town, swim underwater, cast a breathe water spell, and wait for your pursuers to die

no further comment required.
#1720 - NPC lookAt controller. Not gameplay mechanics relevant but it plays a big role in making the NPCs feel "alive", just like facial animations/eye blinking did.
Can't go without #1330 - Cliff Racers can't hit the player. It's not Morrowind

That's just my personal list of very important to fix features/issues, and I most likely forgot some in the back of my head. I'm sure others also have their own list and therefore faster releases = everybody wins.

And also:

optimisation

Yeah, exactly.

[psi29a]

Release early, release often.

You'll not see me complaining on the packaging side.

Debian will always lag a bit if there are suddenly new packages/bits, but for normal releases... those go up pretty fast.

We'll also not have to wait on videos for releases either, which is both a blessing (speedy releases) and a curse (they were AWESOME).

Also... why are we tying OpenCS to a release of OpenMW? I'm getting the feeling that they are two different 'products' with their own tempo. If OpenMW (or OpenCS) is ready for a release, why should we be held up by the other?