We have:
- game design
- physics
- 2d
- 3d
- networking
…but no AI. I’m going to post the question I had that should have gone in an AI category into “tools” instead :(, but I think an AI one would be a good idea too.
This has been requested before, can’t remember what the logic of not adding it was…
I agree tho, it would be a good section…
Kev
Didn’t we use to have it in the previous forum incarnation? Or am I just confused with gamedev.net?
Cas 
+1
This sounds like a good proposal, because AI fits nowhere exactly. Although there will probably be not much activity it’s still good to have it in a central place.
Hmm… content wise (not headline wise) AI fits most into… the physics board. Yes, I know… it sounds a bit odd. But just think about it (in an abstracted manner).
Maybe it would be a good idea to rename “physics” into “logic”?
Well, it makes sense doesn’t it?
@cas
Hmyea we had an AI board before… I’d almost forgotten it.
We seem to be perpetually missing a game logic area, just for general gameplay centered questions. Which kinda covers AI and physics, but is more general. But I think this place has too many sub-forums as it is already. (lots about different technologies and APIs, none on gameplay, does that say something?)
I’m not sure I buy a direct connection between AI and physics or gameplay… but I agree that there are too many forums… and yet the java.net mailing lists go unused… I think we need to strike a balance.
I’m not sure I buy a direct connection between AI and physics
Heh therefore I said “abstract” 
Take several factors, compute around with a common algo or a self written one and update accordingly.
For AI that could mean: check health and ammo, decide if you want to attack/go for health/go for ammo and update the goal.
For physics that could mean: get the forces, apply em to your physics model and update the positions.
Both things need a high level of abstraction, heavy math and alot of dirty cheats
I also agree that there are already a bit too much sub forums, therefore I suggested to merge it with the other sub forum instead of creating a new one.
We have already a “gameplay/design” forum… so… I won’t call the thing orangy talked about “gameplay”. It’s hum… “control flow”?
So maybe…
Game Logic
Discuss game physics, AI and advanced control flow here
Something along that lines… oh and “advanced” because very very simple flow logic like a loop with a 5 case switch doesnt really belong there.
[quote]>For physics that could mean: get the forces, apply em to your physics model and update the positions.
Both things need a high level of abstraction, heavy math and alot of dirty cheats
[/quote]
Well I don’t think ODE is doing a lot of dirty cheats – it is more pure math - like 3D. There isn’t really a lot to discuss about how to such logic in the game - there is usually very few ways that make any sense at all. Most of the Physics discussion is about the ODE APIs for instance… not “how do you accelerate an object?” or “how do you update the position of your sprites?”
[quote]Game Logic
Discuss game physics, AI and advanced control flow here
[/quote]
I agree that AI and general ‘advanced’ control flow can fit together in some sort of ‘Game Logic’ forum for discussing algorithms. As you say, we have a “Game Play” forum already for the non-algorithmic aspects of the game.
Maybe Xith, LWJGL, and ODE APIs have ‘calmed down’ enough… that is they have less traffic… that they can share a “3rd Party APIs” forum all together?
Maybe some of the forums that get little to no traffic can be merged into the general/suggestions areas? (Just update descriptions so people know where to post.)
[quote]Maybe Xith, LWJGL, and ODE APIs have ‘calmed down’ enough… that is they have less traffic… that they can share a “3rd Party APIs” forum all together?
[/quote]
Eh? While the less popular Jogl/Joal/JInput keep their own forums just because they’re semi official? Thats a bit off.
Personally I’d remove Performance Tuning. Yes it gets a lot of traffic, but can’t topics be posted under whatever API they’re discussing? Why should I hunt in other forums just to find out sneeky-fast Java2D (etc) methods? I’m also not sure whether Networking and Online Game Development are really different enough to be two separate areas. Likewise with General Game Discussion and Game News.
Games cheats always - 3D is also just a big fat cheat. Take a look at ray tracing… it’s so incredible slow, despite the fact that it’s totally simplyfied.
Just take jumping in games (eg a first person shooter). It’s simply kicking a up force into that object - the object isn’t pushing itself away from the ground. We won’t calculate the forces wich are applied to the bones, muscles, socks, shoes, the ground, the hair (there’s air heh)… also we won’t be slowed down by dust particles in the air. It’s simplified down to something wich is just enough for our needs.
Or take the bots of Q3. They have an “awareness” box if you shoot into it they know instantly were you are and (haha) if you are from the other team. For the player it looks like the bot heard him, but infact he didn’t - the player just pushed the “I’m here”-button
Physics and AI don’t need to be absolutly realistic - just believable. The majority of CPU cycles is usually needed for the graphics, therefore you just have to find an acceptale compromise.
[quote]I’m also not sure whether Networking and Online Game Development are really different enough to be two separate areas. Likewise with General Game Discussion and Game News.
[/quote]
+1 for merging em
You can use path tracing with Monte-Carlo-Raytracing, if you want to solve the rendering equation completely, but even if you do this you still haven’t modelled the nature perfectly. It will probably scare away every gamer, if you need some minutes to render a single frame.
Xith3D is still one of the most active forums (remember that it’s only about half a year old) and this probably won’t change very soon. We even thought about splitting up the Xith3D forum into smaller pieces, if the traffic increases. I don’t think merging Xith3D and LWJGL is a good idea (LWJGL has its own forums anyway).
I’ve already renderd stuff wich took more than 16hours… just a single 1600x1200 image. That’s ~0.00000017fps :o
But I wasn’t really satisfied with the result. There was still some visible edges despite the fact that I used insane antialiasing settings. Maybe I should have used something like 32bit grayscale instead of 24bit rgb and convert it down at the end (to 8bit grayscale).
[] I agree there are too many forums/cats
[] Given ChrisM’s prior statements on the hassle of re-arranging forums, I suspected rearrangement will be hard to get done, but I thought adding one new cat might be easy enough that it would happen sooner rather than later
[] As the original proposer, I’d be happy with the proposed “game logic” if it explicitly mentions AI in the description, assuming there’s also some other slight clearing-up of the cats.
[] Suggest, instead of calling it “game logic”, call it “game logic and architecture” (or possibly “…and structure”), so that it’s clear if you want advice on structuring your code, it’s the place to look (…and then I’d suggest we put a “standard game-loop” post there and make it sticky! :))
[] As moderator of Networking, and supposedly moderator of Online Game Design (ChrisM said he’d do it, but never got around to it), I’ve often asked what the difference is supposed to be. I’m still not sure :). However, my current interpretation is that one is really low-level stuff, using the networking API’s, etc (which is sufficiently complex and cross-genre to deserve it’s own cat), and the other is high-level stuff, like the design challenges of multiplayer games (cheating, Client/Server vs. PeerToPeer, synchronisation of clusters, etc etc).
[] …and, like with “gameplay”, I would slightly prefer not to have a forum on such high-level things here - there are much much more appropriate places for such discussion, which is entirely independent of programming language.
[] I argued against gameplay all those months ago. Although I didn’t mind, I was afraid it was probably better covered in places like flipcode etc. Looking at that cat now, it seems to have become “game implementation”, mostly topics on what programming designs to use for a game.
[] I suspect game-design is probably well-covered right now in the “your games here” cat: most of the posts (if you exclude bug reports) in that cat tend to be feedback, suggestions, and random musings on game-design
PS And all you raytracing wimps :P, I was doing 1280x1024 renders with funny shaped pieces of glass back in 1992. I used to laugh, smugly, at the poor folk using 286’s for their rendering, whilst I had my shiny new 486 with FLOATING POINT CO-PROCESSOR! Wow, hardware floating-point!. You have nothing to complain about ;).
Actually, RT is pretty fast these days - unfortunately the major supplier (note: …or they used to be the major one; probably not any more: they’ve faffed about so much they surely have been superceded by now…) of hardware RT 3D cards never quite realised that 3D graphics cards for consumers was a big market (nb: the company I’m thinking of has been sitting on tech that for the last 7 years or so was better than anything nV and ATi have produced; but they have a history of making bad strategic marketing decisions :(, and major board-level squabbles over strategy etc ).
[quote]However, my current interpretation is that one is really low-level stuff, using the networking API’s, etc (which is sufficiently complex and cross-genre to deserve it’s own cat), and the other is high-level stuff, like the design challenges of multiplayer games (cheating, Client/Server vs. PeerToPeer, synchronisation of clusters, etc etc).
[/quote]
While the distiction makes sense to me, a look at the current set of topics doesn’t really show that happening. In fact most seem to be low-level stuff or general game logic / architecture. How about ditching the online development forum and adding the logic /architecture one?
Ok, yeah, many recent ones even seem to be in the wrong place entirely :).
Anyway, either way I reckon that “online dev” is not needed :). And, gettting rid of it will knock “assign moderator status for Online Game Dev” off Chris’s todo-list, which I’m sure he’ll appreciate ;D.
I agree with most of the above… and even disagree with myself
I think maybe the core APIs should be lumped together into one ‘Core GTG APIs’ forum… the traffic is low enough (specially in the JOAL and JInput forums).
I think maybe discussion of the development of 3rd party APIS, as opposed to the discussion of development WITH 3rd Party APIs could be moved to mailing lists for such projects. Reducing traffic in the Xith and LWJGL forums.
I know I’ve finally stopped scanning every thread and now resort to the ‘mark all as read’ button.
Anyway +1 for forum cleanup & consolidation and a Logic/AI/Algorithm/Structure section of some sort.
Yeah you definatlyneed an AI forum. And id definately cant be anything to do with physics.
The stuff the AI forum would be talking about is things like path finding using a* algorithms and stuff.
Physics is talking about mechanics. Plus the physics forum here is really about using a 3rd party physics engine, not how do I apply a force here type of thing. The discussions that take place are nothing like what would take place in an AI forum
I throw in my support of a Game Logic board as well.
I also want to throw in my 2 cents regarding the 3rd party APIs. I personally think it opens a can of worms having things set up the way they are. The APIs on the board now appear Sun “supported” if you will, while those that don’t are not quite as official. I’ve long supported keeping jME seperate and maintain it’s own forums, etc. Giving some their own forums but not others causes those without the forums to lose a bit of legitimacy. I’ve e-mail ChrisM requesting a jME forum at the behest of some jME users, but have not received any replies. As such, I’d much prefer that 3rd part APIs be a single forum, so that anything can be discussed without any “Sun Supported bias”.