wraitii

I'm opening this up after reading some more discussion about farms and obstruction sizes. It seems like the consensus in the team is that farms should be bigger, so they're harder to defend. We also discussed making farms depend on terrain and have more profitable and less profitable terrain for farming (see this). I believe the fundamental issue with our farms is that they are large buildings with an obstruction. Making them larger means you will be able to build them in even fewer places, which is annoying. So what's the solution for organic-looking farmlands? Make farms smaller, and limit them to one worker. If our farms were, say, 2x2 tiles, we could plot many around and it would be able to look pretty good (particularly if the decals extend a bit), the fields would actually look occupied by workers (instead of just going around), and it would still take up a lot more size. Then we can add fancyness such as irrigation auras or stuffs like that if we want.

We had a good discussion about the fact the farms should be way bigger back then, but never tried it...

I feel like this is getting metaphysical and the arguments could be said of anything old and new with very little rephrase. Regarding the JIT issue, we mostly discovered that it was a JIT issue, so it would likely be much easier to debug such issues this time (others of new kinds could, it is true, arise). What we have already discovered is that the new Spidermonkey is stricter about lax syntax and things like that, which are sources of weird, unpredictable, and hard to debug issues. We also know that it supports multi-threading to a greater extent. And regarding performance, there are several stuffs beyond just compilation magic that could help us, such as new objects (Maps and Sets) which possibly could speed the AI a lot. On 1.8.5, I do not believe the AI will ever be efficient enough. GC is quite likely to get faster, particularly if it can be fragmented (which ought to be able with the newer spidermonkeys). If anything, this should be a separate branch on Git. Somewhere you can just try it easily at some point, to see how some changes react. If we're going to get economic, the opportunity cost of not trying the upgrade is just too high.

I'll give it a look. There's an issue currently, the AI gets stuck on farms, it's because of this morning's changes, I'll fix it tomorrow.

Rodmar: interesting ideas, I hadn't really though of many ways to extend this but that could be nice. Like a "stone age" gameplay that tries to still be fun without all the fancyness. Ludo: Thanks

I agree with Micket that it's too blocky so far, but I encourage you to do your own too, as we do want variations. if you know how to, perhaps create a very high-poly model with sculpt or textures to achieve a grainy "old granite" effect for the normal textures.

Could you give me some input on how early an "early rush" would be to you? Also with how many units did you attack?

Gameboy: I haven't committed the fix yet.

Gameboy: That's now fixed Lion: thanks for those, I don't think they're fixed yet, I'll look into it.

greenknight: did you say you saved and reloaded before trying? Because that would break the AI in many ways and is expected not to work too well. On the other hand, I haven't noticed the AI getting stuck on Alpine Valleys. It started really slowly on Deccan plains but that has been fixed, so now it should work mostly properly. Still, it's far from perfect anyhow.

I've given a look at this the other day and have been completely unable to understand what happens. Collada seems to load the skeleton properly, so the issue is with the model. But importing/re-exporting in Blender fails. And there has been no obvious changes that would have broken the boar.

Those errors should be fixed now, thanks for reporting them.

I agree with Yves that the GC improvement in the future seem far from trivial to me. GC is slow in 0 A.D. and leads to a bunch of other issues, so that would be very nice, and if only for that, the upgrade is worthwhile.

I've fixed the first one, I'll fix those two tomorrow, thanks for reporting them.

Hi Gameboy, I'd be interested in having the commands.txt of that error (or specific steps to reproduce it). I understand what happens (and it's now fixed) but I'd like to see the circumstance in which it happened.

fabio: those warnings are stuff I should have fixed and forgot to before committing, I'll do it now. greenknight32: yes that's a known issue, rushing the AI early enough basically means killing it right now.

I see. Then it's probably better to have ESR 24 as an intermediate, yes.

I've just committed some WIP changes to aegis, which should work with Yves' changes but haven't been extensively debugged, so please report errors that could come up and weird behavior. It should fix an error in the economy (the AI didn't actually care about full resources) so that should improve the econ a fair bit. The defense manager is being rewritten from scratch too so that could change, right now I think it's a little worse than it was before, but it's all WIP. I've removed older AIs too.

I can't say much about the first two options, I believe we relyon our spidermonkey on OSX so I don't really care. Are there big changes in ESR 31 compared to ESR 24? (I'm assuming that ESR is extended support release). Also, I find it amazing that the newer spidermonkey is always slower than current SVN. It's so counter-intuitive.

Well trimmed they're supposed to look like this which isn't extraordinarily different.

We don't have incremental updates yet (mostly because most developers use SVN so that's a non-issue for us), but we intend to implement something like that someday.

It's just that I know Blender (I can understand not wanting to use it if you're not used to it though) and there are easy settings to set environmental lighting and ambient occlusion, which looks a lot more natural? 3DS Max might have those too, no idea.

Actually, running/charging wil probably not be done with stamina, but it's not completely decided yet.

stanislas: no problem, it's just that if you were in Blender I could have given you a few pointers on how to improve it quickly.

What software are you rendering in? I don't like your lighting setting, I find it quite unrepresentative of how the building will look in-game. I think your wooden planks on the side look weird. Too regular when I'd expect them not to be nearly that much. Also there's no apparent fixation for them, which seems a little off to me.