John Carmack's Keynote Highlights

[if76]

John Carmack, the man who needs no introduction apparently had quite a bit to say at this year's QuakeCon.

Read the short version (http://www.gamasutra.com/php-bin/news_index.php?story=6212), the long version (http://www.gamespy.com/articles/641/641662p1.html), the unedited version (http://www.ga-forum.com/showthread.php?t=59227), or just download the footage and watch it (http://www.filerush.com/download.php?target=QuakeConCarmackKeynote.zip).


There used to be a saying back in the early days of 3D rendering where someone quipped, "Reality is 80 million polygons per second." It was a great catchphrase at the time because during those days, it was an obscene number -- ridiculously beyond the technology of the day. But nowadays, Carmack points out, even cheap console systems can crank out that kind of performance. More importantly, today's technology can render that many polygons in realtime with tons more features than was ever available before, such as programmability or multiple passes. And the high-end keeps getting higher.

At this level, are graphics all that important? Carmack says that sometimes he gets static from other game developers, who say (only half-jokingly) that id Software has raised the graphical bar so high that other companies have to scramble and dedicate tons of resources to catch up. Once consumers see that level of graphics, they expect it from every title. Carmack says he's sympathetic to that, he understands where they're coming from. But at the same time, he made it clear that he really sees huge value in what advanced graphics bring to the table. People can talk about how gameplay is king, but he asserts that technology allows you to make a really memorable experience. "Presentation really is important," he explains. He admits it would be nice not to have to do tons of motion-capture and cinematic stuff with every new game released, but argues that that extra layer of fidelity makes games better.

People may have fond memories of simpler gaming eras, but as much fun as Quake I was for its time, he asserts that it just doesn't have the presence or impact of modern games. "The golden age is right now!" he argues.


Clearly what the game industry needs is a better type of normal mapping.

[Varsity]

That's out of context. He is talking about a full, physically simulated environment with deformation and all of that stuff, not something like HL2.

[MasterEvilAce]

I could agree with that, a more constant framerate is always good. Physics on a tv doesn't please me much.. and wtf is with physics in some of these games? A guy gets shot and his body just instantly... limp.. rolls around on floor for an entire minute before stopping in an almost impossible position. While I like how the Source engine handles a lot of physics.. they slack on so much, as well.

Who the hell are the ones that are yelling, "Physics!" and "Graphics!" all the damn time?
A good game concept can go horribly wrong because it can't "compete" with the standards that are getting set, so it's harder for a new company to start up and enter the mix.

[Steve_Erhardt]

Physics are nice, but if you have to choose true-life physics over graphics or gameplay, I'm perfectly okay with scripted physics over dynamic physics.

[Conner Dain]

I find it amusing to see anyone being critical of John Carmack. Cheap shots are soooo easy. But I would stack his life accomplishments against anyone else in the industry. Few have done as much, let alone more for PC gaming. Like to play FPS games online? Thank JC. He invented the event prediction system that practically all FPS online games use in one way or another. Like multiplayer FPS games? He invented that as well. I seriously doubt he loses any sleep because DOOM III wasn't the perfect game.

[bean19]

Wow. Bad reporting.

[Mister Pie]

I personally prefer realistic physics at the expense of some graphics. It makes the game so much more believable.

Of course, I'm not saying I'd prefer like Duke Nukem 3D style graphics with the most advanced physics engine in the world, but a small hit to graphics would be alright by me.

[thecrazyd]

I find it amusing to see anyone being critical of John Carmack. Cheap shots are soooo easy. But I would stack his life accomplishments against anyone else in the industry. Few have done as much, let alone more for PC gaming. Like to play FPS games online? Thank JC. He invented the event prediction system that practically all FPS online games use in one way or another. Like multiplayer FPS games? He invented that as well. I seriously doubt he loses any sleep because DOOM III wasn't the perfect game.
So what? He created the FPS. Great. I give him props for that. I do not give him props for his desire to not move foward at all.

[Last of the Red Hot Mamas]

It doesn't seem to me like he's discounting physics altogether. He seems to be more concerned about portability to consoles than anything else, and in that respect he's probably right -- without any sort of dedicated physics processing on the next-generation consoles, you're not going to see a whole lot of developers taking advantage of the technology. It's sad that innovation in the PC world is now being hamstrung by the limitations of the console world, but that's just the way the market is. I doubt physics processing will take off until it's incorporated into consoles, so we're talking at least another five or six years (discounting the very unlikely possibility that the Revolution will include some kind of PPU).

[Doomsday]

Clearly what the game industry needs is more crates to knock over

[Cupelix]

That's out of context. He is talking about a full, physically simulated environment with deformation and all of that stuff, not something like HL2. Do you know its out of context because you heard the whole speech? The article does sort of gloss over what level of physics integration he is talking about.

[Zanzibar]

He's secretly hoping for a 'monster closet' chip. {/rimshot}

[RMan]

The article does sort of gloss over what level of physics integration he is talking about.
This is true, and without that it's pretty hard to understand if he's overly against physics or not. IOW, he may be seeing the possibility of moderate physics implementation dropping framerate from from 60 to 30 fps, in which case it’s arguably not worth it. Of course, a really good implementation of physics would make this framerate hit worthwhile, IMO (although I’ve not seen a game yet that falls into that category, but physics is relatively new). It’s kindof an easy statement for him to make though, since the next generation won’t have hardware assisted physics there’s a clear limit to what can be accomplished in the physics without a drastic performance hit. Beyond next generation, though, I expect physics will set the world on fire FAR more than ANY graphics improvement will.

[Borys]

Did Carmack banged your wife? That's a really bad newspost, if76.

[Rafer]

Read the entire transcript here (http://www.ga-forum.com/showthread.php?t=59227).

And while I can sort of see Carmack's point, there's just something really compelling about knocking stuff over. I've had a lot of fun with things like the Novodex demo and Garry's Mod. I would just like to see a proper game based around that stuff, even if the physics take such a processor toll it result in something like Katamari Damacy level graphics.

[XenonCJ]

So what? He created the FPS. Great. I give him props for that. I do not give him props for his desire to not move foward at all.Something tells me John Carmack knows what "moving forward" means more than you do.

[Cupelix]

Beyond next generation, though, I expect physics will set the world on fire FAR more than ANY graphics improvement will. I'm with Carmack in the fact that I don't find integrated physics to really enhance my games that much. The only thing I really want out of a physics engine is for things to fly off shelves when I accidently shoot them in a battle, and for people to crumple realistically when they die (most games have still not gotten a handle of this). When some of these things are done poorly, it actually distracts me more than in aids in suspension of disbelief. If a bad guy falls to the ground and his foot ends up behind his head...

Anyway, I think having fully integrated physics will actually be a pain in the ass for developers, because sometimes characters need to move in ways that aren't physically possible - the speed and height of jumping comes to mind.

Again, it all depends on what makes a game fun. Moving planks and boxes around in HL2 so I could jump from one side of a room without touching hazardous substance X was not fun. Yet, if the game didn't have a physics engine, that "puzzle" wouldn't be there in the first place.

[emperordahc]

Perhaps not 100% realistic physics, but I'd sure like something better than Doom3's propriatary physics. Havoc is pretty good in anything, like Halo2.

[Liquidize105]

Yeah, I read gamespy's report. He was doing no such thing.

I'm moving this.

[Bone]

I have an immense respect for Carmack, he really did pioneer so much of what we take for granted today.

At the same time, it's been clear to me as they moved from Quake 1 to Doom 3 that Carmack is focused more on technology than gameplay. As such, I take his technical comments on physics a lot more seriously than his gameplay-oriented comments on physics. He may not be convinced that physics are important enough to warrant their own chip, but there are creative possibilities with physics that I guarantee we will be taking for granted in 10 years.

[ÜberJumper]

Yeah, this is not main page material, especially with a topic line that was so out of context.

[KNOTE]

Clearly what the game industry needs is more crates to knock over

In the future, we will need some space crates to knock over too.

[Mr.Green]

I don't want to be needlessly rude but that's sensationalist tabloid shit from a bull.

[Atorak]

Carmack was not in charge of making the gameplay of Doom III fun, he was primarily focused on the engine development. Sure, you can look at newer, possibly more accessible engines, such as HL2, U3, etc, and when you do that, it becomes "easier" to point the finger at Doom III's "flaws". But, the fact of the matter is, Carmack made a hell of an engine, that ran fantastically on every system you threw at it.

[begin side rant]

I really enjoyed Doom III. I was literally jumping out of my seat all the time, it was creepy as hell, and it had a bunch of memorable moments. Frankly, that was enough for me. But, I wasn't playing with the lights on, next to my parent's bedroom, with 2 Logitech speakers pumping out a massive 2.5W/channel either. Doom III required your full attention, it required a dark room, with loud speakers, a kicking sub, and a good PC. Since I had all those elements, I didn't really mind all the "monster closets". I was so freaked out most of the time, that I was basically just swinging around with my shotgun, pumping shots into the dark, hoping to silence the huge beast screaming at me.

Anyway, I love the fact that new games like Fear can build on what Doom III started, but I won't discredit an original title.

[bean19]

Read the entire transcript here (http://www.ga-forum.com/showthread.php?t=59227).

And while I can sort of see Carmack's point, there's just something really compelling about knocking stuff over. I've had a lot of fun with things like the Novodex demo and Garry's Mod. I would just like to see a proper game based around that stuff, even if the physics take such a processor toll it result in something like Katamari Damacy level graphics.

Great link.

Just for those who aren't interested in reading that whole long transcript, I'll quote the part of John Carmack's speech regarding physics cards:


From a transcript of John Carmack's remarks:
Ok, stand-alone physics cards. They’ve managed to quote me on the importance of, you know, physics and everything in upcoming games. But I’m not really a proponent of stand-alone physics accelerators. I think it’s going to be really difficult to actually integrate that with games. What you’ll end up getting out of those, the bottom line, is they’re going to pay a number of developers to add support for this hardware and it’s going to mean fancy smoke and water, and maybe waving grass on there. You’re not going to get a game which is radically changed on this. And that was one of the things again why graphics acceleration has been the most successful kind of parallel processing approach. It’s been a highly pipelined approach that had a fallback.

You know in Quake, GLQuake, and Quake 2 timeframe where we had our CPU side stuff, and the graphics accelerator made it look better and run faster.

Now the physics accelerators have a bit of an issue there where if you go ahead and design in these physics effects, the puffy smoke balls, and the grass and all that, you can have a fallback where have a hundred of these on the CPU and a thousand of them if you’re running on the physics accelerator.

Once of the problems though it’s likely to actually decrease your wall clock execution performance, and this is one of the real issues with all sorts of parallel programming is that it’s often easy to scale the problem to get higher throughput, but it often decreases your actual wall clock performance, because of inefficiencies with dealing with it. And that’s one of the classical supercomputer sales lines that you can quote these incredibly high numbers on some application, but you have to look really close to see that they scaled the problem.

Where usually people when you think of acceleration you want to think “it does what I do only better and faster” and a lot of cases in parallel applications you get something where “well, it does what I do, it’s better, but might actually be a little bit slower”, and this was one of the real problems we had with the first generation of graphics accelerators until 3DFX really got things going with Voodoo.

A lot of the early graphics accelerators you’d take the games, they would run on there, and they would have better filtering and higher resolution, so in many cases they’d look better, but they were actually slower -- in some cases significantly slower than the software engines at the time. It was only when you got to the Voodoo cards that actually it looks better in every respect and it’s actually also faster than the software rasterizer version that they became a clear win.

So I have concerns about the physics accelerator’s utility. It’s the type of thing where they may be fun to buy for their demos, it might be cool, and there will be some neat stuff I guarantee it. I know there’s some smart people at the company working on it that I’m sure will develop some great stuff, and there will probably be some focused key additions to some important games that do take advantage of it, but I don’t expect it to set the world on fire really.

[Cupelix]

The transcript is a great read, incidently - that is worth having a front page news article about. I would like to note one particular line about why he thinks the way he does: This is neat stuff, but it remains kind of non-core to the game experience. He's focused on what tools and technology he thinks improve the actual game experience. To him, graphics will always be top of the food chain, because games still don't look like the real world. Everything else can be faked on way or another.

[XenonCJ]

Yeah, this is not main page material, especially with a topic line that was so out of context.How's that? I thought this was a gaming news site? The future or lack-there-of of gaming hardware certainly qualifies.

[bean19]

How's that? I thought this was a gaming news site? The future or lack-there-of of gaming hardware certainly qualifies.

Xenon - They've fixed it now, but the original news posts was "Carmack badmouths physics cards" (paraphrased as I don't remember it), and the story that was linked was highlights of all the things he spoke about. . . not just physics cards.

However, even the story linked does not give a full accounting of Carmack's statements and is misrepresentative in the particulars.

For a hardcore gaming news site like EA, one would expect the link that Rafer created more than the link to the slightly yellow gamasutra story.

[Zanzibar]

Hey, does it bother anyone else the 'ragdoll' fad that's running throughout videogames nowadays? As soon as a character is 'killed', they instantaneously become floppy sacks of bones that have no ability to move.

I don't know what it is, but it just looks stupid. And it's fuckin' everywhere. I miss death animations, where the character drops to his knees and falls over. THEN, after that, you can activate the ragdoll and allow the body to fall down stairs etc.

[Liquidize105]

Yeah, I fixed it up somewhat.

And as for rag dolls, well you gotta start crawling before you'll be walking and running. Shooting someone head-on between the eyes and have the fella fall forward and die, now that really "looks stupid." The whole "going limp" thing is a tech phase, though I agree havoc is overrated for the time being.

[Bone]

After reading the transcript it's even more clear to me that this is Carmack saying physics gameplay basically doesn't excite him. "This is neat stuff, but it remains kind of non-core to the game experience." I disagree: it's only non-core to the experience because we haven't been able to take advantage of it and make games where it IS core to the experience.

I also don't think it's wise to wait until visual reality is "done" before moving on to physics. We're at a point where things already look real enough that it seems unrealistic when everything is nailed down to the floor. It's not just about prettier water and grass blades, as Carmack says in the transcript. I think for some games, the game world needs to start behaving as realistically as it looks and sounds.

[ÜberJumper]

Yeah, the post works better now that Liquid's tuned it. Before it was nothing but a big potshot at John Carmack.

[Zanzibar]

And as for rag dolls, well you gotta start crawling before you'll be walking and running. Shooting someone head-on between the eyes and have the fella fall forward and die, now that really "looks stupid."

I agree entirely. But the technology has existed forever to do hit checks by the vector and adjust which anim to play accordingly. Maybe I saw too many action pictures, but when I land a grenade in a group of guys, I want to see them flailing in midair before they hit the turf. The instant transformation from 'attacking enemy' to 'lifeless sack' reeks of lack of style and/or laziness. Halo 2 was terrible at it - compare the satisfying death anims of the Covenant forces in Halo 1 with Halo 2's style-less, floaty deaths and you'll see an example of why ragdoll is overrated.

[Bone]

Ragdoll isn't overrated, it just seems like we don't have the horsepower to implement it properly.

To do what Zanzibar wants, you need to be able to incorporate AI movements into the ragdoll- so that when you get shot in the arm, your body recoils but your other hand may reach up to the wound. Or in the case above, the grenade throws the enemy using physics, but they also flail about using pre-animated "fear/flailing" animations. For this, you have to be able to blend animations, applying physics to an existing animation. All of which costs precious cycles, and why it seems like we do need a dedicated physics processor.

[bean19]

I want another SOF game. . . please make more enemies who run screaming after you sever their limbs.

There were definitely a lot more clever death animations before developers moved to ragdoll (which can be cool. . . but mostly after huge explosions). Maybe just have ragdoll for the replacement for "gibs" and do smaller firearms deaths with interesting death animations? Ragdoll and clever death animations aren't opposing goals after all.

[Heretic Machine]

It doesn't seem to me like he's discounting physics altogether. He seems to be more concerned about portability to consoles than anything else, and in that respect he's probably right -- without any sort of dedicated physics processing on the next-generation consoles, you're not going to see a whole lot of developers taking advantage of the technology. It's sad that innovation in the PC world is now being hamstrung by the limitations of the console world, but that's just the way the market is. I doubt physics processing will take off until it's incorporated into consoles, so we're talking at least another five or six years (discounting the very unlikely possibility that the Revolution will include some kind of PPU).

I hate to break this to you, but a lot of us PC gamers find the idea of a PPU to be dubious as well. I don't plan on buying one anytime soon. To me, it seems a lot like somthing to boost hardware sales...

[RMan]

Ragdoll isn't overrated, it just seems like we don't have the horsepower to implement it properly.
What you have to do is have animations, or AI driven reactions to stimuli, then push the body parts towards the intended body part destinations using the physics engine. Although this does take more processing power, the real reason we haven’t seen it is because of the difficulty of implementing and tweaking it, we most assuredly have the processing power right now. Having instantly set animations (classic animations) or totally physics driven (ragdoll animation) are both very easy in comparison to the more intelligently driven physics/animation based hybrid.

[Borys]

LiquidSnake to the rescue! That's a much, much better reporting job, thanks LiquidizeOneOFive!

[jeffool]

Wow, I want to see the original post now. :D

But I agree that physics are fudged well-enough to the point that I don't see how perfect physics would be an 'amazing improvement!'. Sure it'll be good, but now 'omgwtfbbq'-good. And the marriage of ragdoll with traditional animation seems to be a fine idea for most non-game-affecting animations.

/edit: As a side note, everyone who's DLed the video from FileRush, keep those torrents open and help a man out! This 15KB/s stuff is ridiculous.

[Zanzibar]

I want another SOF game. . . please make more enemies who run screaming after you sever their limbs.

There were definitely a lot more clever death animations before developers moved to ragdoll (which can be cool. . . but mostly after huge explosions). Maybe just have ragdoll for the replacement for "gibs" and do smaller firearms deaths with interesting death animations? Ragdoll and clever death animations aren't opposing goals after all.

Right. Halo 2 is a terrific example. If you shoot an enemy in the arm for its final hit with a pistol, it still becomes ragdoll floppy for its entire body. If it cried in agony and passed out, hit the floor and died - as in Halo 1 - it wouldn't look like 'ragdoll physics took over because it's instantly dead.'

Shoot a Halo 2 Jackal in the head with a sniper rifle and watch the floaty physics in action. Gah. Just laziness - task a few animators to have satisfying death anims and have ragdoll take over either when the anim is interrupted by collision with level mesh, or when the animation is done (thus allowing bodies to fall down stairs etc.)

I know, I know, I'm picky. But seriously, has anyone else noticed this?

[bean19]

One of the funniest things in Hitman 2 is when you kill anything on steps. . . They'll slide down the entire steps. Boy did I want a co-op mode. My friend and I wanted to use bodies for a "slinky" race.

[snubber]

Well hopefully we'll see something like this product in games:

http://www.naturalmotion.com/pages/products.htm

It's not just lame ragdoll, but AI-driven animation generated on the fly given the circumstances. Hopefully one day a game will have something similar, so if a body is blown to the ground or something, you'll see not only realistic constraints (knees that don't go backwards) but also realistic motions, like the guy putting his arms out to instictively break his fall.

[Last of the Red Hot Mamas]

I hate to break this to you, but a lot of us PC gamers find the idea of a PPU to be dubious as well. I don't plan on buying one anytime soon. To me, it seems a lot like somthing to boost hardware sales...

That seems like a given; I can't imagine anyone releasing a piece of hardware with the deliberate intention of not boosting hardware sales. But then I don't plan on buying one anytime soon because developers aren't going to use them for anything particularly compelling, due largely to the scalability and consolization issues -- if you have to make your game run on PPU-equipped systems and non-PPU-equipped systems and consoles, you can't use the PPU to do anything critical for gameplay, so there's no compelling reason to have a PPU. That strikes me as the gist of Carmack's remarks.

[bKangy]

I'm really not that bothered about physics in gameplay compared to say... AI. The physics in HL2 were fine, to be honest. Ragdolls are completely overrated, but if working on some improvements to that instead of more advanced, realistic AI isn' really want I want.

I'd rather AI actually noticing and reacting to the fact you just shot their friend dead, rather than that friend going through a full physics death animation.

[ElectricMonk]

hmm i think if you read between the lines in these complaints you can actually see what gamers _really_ want.

physics is the new coloured lighting. there was a time when ragdoll deaths were new and awesome and you purposely tried to get the bodies into awkward death positions.

there is a way to combine ragdoll effects (IK) with animations so you can get the best of both worlds so we'll probably see that in the future.

[Tennistoad]

Wow to read between the lines it sounds like Carmack is mad that AGEIA physX went to epic with their new unreal engine instead of going to him like all the video card makers do. Sorry but you really do need good physics more than graphics for an awesome racing game.. nascar season 2k3 proved that along with papyrus gp. Without the physics some games just don't really work for what they are trying to do.

But on the other hand, I could use a really good flashlight simulator!!

[RMan]

physics is the new coloured lighting. there was a time when ragdoll deaths were new and awesome and you purposely tried to get the bodies into awkward death positions.
Now, now, let's not make physics synonymous with one relatively small use for it (ragdoll). Physics is very new, and not much has been done with it yet, good implementations will allow new gameplay experiences, which colored lighting, or any graphical improvement since we went to 3D can not offer.

[RMan]

Wow to read between the lines it sounds like Carmack is mad that AGEIA physX went to epic with their new unreal engine instead of going to him like all the video card makers do.
It's kindof interesting, but I think id pretty much screwed themselves by going with their own, proprietary technology on this one (only obviously a bad move in retrospect, though). This is speculation, but I think the main reason Epic went with Novodex is that it was a relatively undervalued system that had no big licensees, and they needed something better than Karma, which didn’t seem to be getting any better. They wouldn’t want to go with Havok (which was really the only other engine worth considering) because they wouldn’t be able to be the clear premiere licensee and thus have greater control over the technology due to Source/Valve using them, as well as Havok being a much stronger player in the market overall. Ageia’s acquisition of Novodex also likely had to do with their being undervalued, since they were really almost as good as Havok but didn’t have the serious pricetag that would have come with joining with Havok to make the hardware. If id had actually looked around a bit they would have seen the physics engine (previously known as Steel) and could have licensed it for a song instead of implementing their own, but without the foresight of the possibility of a hardware physics chip making your own physics engine doesn’t seem like such a bad idea. Anyway, I think Epic’s use of it was more good fortune and how things fell in place than it was a well thought out strategic move.

[Conner Dain]

IF76, I see you've edited your cheap shot at JM. How did your words taste? (Seeing that you had to eat them.)

[Mr.Green]

Well hopefully we'll see something like this product in games:

http://www.naturalmotion.com/pages/products.htm

It's not just lame ragdoll, but AI-driven animation generated on the fly given the circumstances. Hopefully one day a game will have something similar, so if a body is blown to the ground or something, you'll see not only realistic constraints (knees that don't go backwards) but also realistic motions, like the guy putting his arms out to instictively break his fall.
Wow that's awesome! :eek:

Great find snubber.

[maja]

Ragdoll physics is just rediculously silly when used in these FPS games. Anyone remember Sodlier of Fortune 2 and its G.O.U.L.2 technology? Man, that game had some awesome death animations. Yes, they where flawed but it was sweet blowing people up in that game if only to watch them die in agony. The biggest flaw was when they would fall right through a wall.

So maybe we need a combination of death animations with ragdoll physics. Like let the animation do its thing and then once the person is "dead" kick in the ragdoll stuff to prevent bodies going through walls or bodies falling the wrong way. Someone needs to make this work!

[Mason]

Gameplay.

AI.

How ridiculous is it that everyone is expected to spend more and more money improving the superficial aspects of their game instead of tackling any of the really interesting problems?

Enemies in action games still stand around stationary waiting for PlayerController to stroll by. Then they employ tactics that wouldn't keep them alive in a game of dodgeball with third graders. Making stupid enemies more and more detailed in appearance is a dead end road for gaming.

I don't get how people can complain about minor flaws in graphics and animation breaking immersion, when the NPCs in whatever game they're playing act nothing like real human beings. All the bump mapping and inverse kinematics in the world can't provide realism if you can't produce human beings that can navigate and interact with their environment in realistic, unscripted ways.

Kudos, we won on graphics. Image quality is rapidly approaching Good Enough, after which point it is only a waste of time and money to dink with them. Now, let's see some headway on the real, tricky problems.

[bean19]

I don't get how people can complain about minor flaws in graphics and animation breaking immersion, when the NPCs in whatever game they're playing act nothing like real human beings. All the bump mapping and inverse kinematics in the world can't provide realism if you can't produce human beings that can navigate and interact with their environment in realistic, unscripted ways.

Kudos, we won on graphics. Image quality is rapidly approaching Good Enough, after which point it is only a waste of time and money to dink with them. Now, let's see some headway on the real, tricky problems.

Well, that's the topic. Maybe we should talk about AI all the time even when it is not the focus of a thread?

Also, AI isn't uniformly terrible. My UT2K4 bots often surprise me with their intelligence. I could name numerous other exceptional AI games, but, you know, it's not on topic.

[Liquidize105]

Name them. UT2k4's the only game I can think of with surprisingly good AI.

In a nutshell, AI is uniformly terrible with the occasional exception.

[if76]

Ok I got a lot of negative responses on my original post (before it was edited to its current form) and I admit to being a little harsh. Perhaps I should have started by saying "with all due respect". That being said Carmack hasn't really moved the genre forward since Quake and many will say that bringing a FPS into full 3d was an inevitable step. The guy is a techincal wizard but he's not a very imaginative game designer. The single player games of Quake, Quake2, and Doom3 are all pretty forgettable and none of these represent much of a departure from Doom2.

It's also pretty clear by the way he gushed over the graphics engine in Doom3 that he's a pretty big graphics whore and I don't think I'm alone in saying that the industry is a little bit too graphics obsessed.

So when a man who carries a lot of clout in the gaming world who hasn't created an innovative gaming experience since the mid-90's comes out and says that physics aren't important, it upsets me.

This is mainly due to the fact that two of my favorite games in recent years Gish and Psi-ops are very physics-based. While one could argue that Half-Life 2's physics were the icing on the cake, in these two games, they were the icing, the filling, and the yeast that made them rise.

I think this next generation of gaming needs more Gish and less Doom 3.

[if76]

IF76, I see you've edited your cheap shot at JM. How did your words taste? (Seeing that you had to eat them.)

I did not edit the post. It was edited by ea staff. I stand by my words and am no where near hungry enough for your acceptance to eat them.

[bean19]

Name them. UT2k4's the only game I can think of with surprisingly good AI.

In a nutshell, AI is uniformly terrible with the occasional exception.

The bots in Counter-Strike are really stupid at low-level settings but are brilliant at high level ones. Which makes sense given their difficulty slider.

Also, I found the AI in Half-Life 2 to be appropriate to the mobs. Squads fought as squads and found cover, and head crab beasts took the fastest route to you they could find. The friendly AI was stupid, so this isn't an infallible example, but I'd say that the AI overall was good.

I also liked the AI in Halo and Halo 2. They knew where larger weapons were and they often zeroed in on MasterCheif as a larger threat than the marines and focused on him (even yelling out things).

Really loved the A.I. in the F.E.A.R. demo. I played through that several times and each time they responded to my different tactics in different (and appropriate ways).

Not all AI is awesome, but it is quite passable in other games. The mobs behaved with some intelligence (use of cover when it is available) in the majority of FPS games I've played. . . SoF 2, NOLF 1 & 2, Rainbow Six.

Only rarely is AI truly horrible. . . Battlefield 2, Star Wars: Battlefront. In Jedi Academy, the AI was actually pretty good as far as battling intelligently. The AI's key flaw was being stupid enough to get near ledges all the time. This was obviously intended in the level design for the singleplayer game (pushing mobs off a speeding air train in Coruscant is awesome and cinematic.) but they shouldn't have been as stupid in multiplayer.

So I'd say passable AI is the norm, with plenty of examples of terrible AI and exceptional AI.

[Roc Ingersol]

PPUs are overrated simply because multithreading is hard.
Because it's inherently hard, physics becomes the first, most natural thing to move to another thread or processor.

As soon as you get dual-core chips (e.g. nextgen consoles, next gen PC CPUs), the physics processor's job just got co-opted by the base hardware.

The 360 and PS3 have PPUs, effectively. Sure, a proper PPU would be more hard-coded to the task, and could push more calcs per second, but at what cost?

Frankly, if multi-core CPUs hit before GPUs did, those probably wouldn't be a big deal either.