Carmack Keynote Info

I wanted to respond to a few of you, so I'll just post this top level.

Those of you who say JC is not taking physics seriously aren't getting his point The thing that allowed adoption of 3D graphics was that is could be pahsed in slowly. Some people had accelerators, some didn't. And at the end of the day, with Quake2, you could play the game the same way whether you had a 3D card or not.

Even when 3D cards were commonplace (like now), if you have a bad card, you just tone down the settings, but the gameplay stays the same. What JC is pointing out is that with a physics accelerator, this isn't possiblre in the same way.

I'll use the same word he used: it doesn't scale as well as 3D acceleration did. When you introduce a new feature that affects gameplay, you have to be very careful that the entire audience that buys the game will have access to that feature. As he already pointed out, if you try to introduce physics to a game in a way that makes it optional, it will end up only affecting parts of the gameplay that are optional - like water, wind, trees, stuff like that. You can't make puzzles centered around having a physics accelerator, because a lot of your audience won't have access to that.

And then you get to the chicken and egg problem: no one will want to pay for a physics accelerator because it doesn't change gameplay. No games will be released that make use of a physics accelerator because they can't afford to market a game ONLY to people that are in the top 10% with respect to hardware.

I hope you see what I'm trying to say: JC wasn't bashing physics as being important, he was pointed out a large barrier to adoption which has to do with a phasing-in process for the hardware and the games. And, as he started out saying, that has to do with the nature of the feature. In the case of 3D cards, it worked. He's arguing that with physics cards, it'll be "iffy" because it'll be harder to phase in.

As usual, I think the guy is spot-on, but I know the shack isn't the place for id fans. =)

I thought that this is what the fpu is for. A good fpu (i.e. not the Pentium 4) can blaze through vector calculations needed for physics engine work easily.

Back at the advent of 3d accelerator cards, weren't there games that you couldn't play without them? And didn't people buy the cards so they could play?

GRAPHICS are peripheral to the core gameplay experience

I really wish he would realise this. I am a graphics addict, I love seeing beautiful games. However I also like games that are fun and non-repetative. Between this comment and his comments on story/plot in games, I have little respect for the man. He is an amazing graphics programmer, but should not be designing games.

I don't understand why they don't just stick these chips on video cards. That would almost ensure their place in the market.

So how many instances of "MN!" this year?

I was disappointed at the lack of graphics discussion that he will use in the new engine. But I guess he wants to save that up for another time.

HL2 basically proved that Physics could be a HUGE part of the game and not something like eyecandy.

I like Carmack's outlook.

He seems to retain the "forest through the trees" perspective.

This is a stupid argument.
If your simulating the real world, then you need to simulate physics. If all people wanted was boring eye candy, they could just play doom3 for the rest of their lives.

Physics arent a gimmicky game play option, there an integral part of a simulation that enhances the immersion into the environment. As long as developers look at it this way people will continue to consider 3d titles as "games" that nerds and little kids play.

Lets face it, some day these titles will replace soap operas as the entertainment of choice for a new generation. They'll be just as real as the room your sitting in today. So why complain about the destination instead of getting up off your programming ass and taking the next step towards the synthesis of the real...

Does Cliffy B ever have a keynote? I'd rather hear what he has to say about physX

There's something ironic about hearing Carmack speak about gameplay over realism.

Does the main link to the notes crash firefox for anyone else?

Speaking of physics level of detail, in the game I'm making I actually do kind of do this; I can flag objects as "insignificant" indicating that they should really be given real calculations if they are visible, otherwise they just get trivial stuff like plain newtonian motion. Additionally, they can either be handled by the physics code itself or by ODE, the physics API I'm using for the fancier stuff.

Of course, the physics in a shootemup are quite peripheral to the main game experience but I think watching a dying boss spiral out of orbit and into a moon along a proper trajectory while belching out fireballs would just look really cool :)

That being said, there's going to be lots of stuff blowing up.

Full transcript: http://www.ga-forum.com/showthread.php?t=59227

doctors should play a recording of this to chronic insomniacs. this could be a miracle cure for them.

Stupid linking included the full stop...

http://www.filerush.com/download.php?target=QuakeConCarmackKeynote.zip

You can all watch this - http://www.filerush.com/download.php?target=QuakeConCarmackKeynote.zip.

To use the chip as something integral to gameplay is going to take a developer who is willing to make the card something necessary in order to play the game as id did with graphics cards when Quake 3 came out. It took three years from when graphics cards were optional enhancements that provided interesting but unnecessary enhancements to lighting and quality to when they were necessary to further develop the game. I wonder if physics will catch on faster than that 10 years after the last development major in gaming technology first emerged now that more gamers are more willing to spend $$$ for a top machine, or if the much longer development timelines will mean that physics is going to take much longer to be implemented.