Carmack leaves Id
polycounter lvl 19
Just seen this posted on JC's twitter feed. End of an era.
http://www.gamasutra.com/view/news/205539/John_Carmack_officially_leaves_id_Software.php?
"In two separate tweets, Carmack (@ID_AA_Carmack) said, "I wanted to remain a technical adviser for Id, but it just didn't work out. Probably for the best, as the divided focus was challenging.
"If they don't want me to talk on stage at Quakecon next year, we'll just have to fill up the lobby like the old days. :-)"
http://www.gamasutra.com/view/news/205539/John_Carmack_officially_leaves_id_Software.php?
"In two separate tweets, Carmack (@ID_AA_Carmack) said, "I wanted to remain a technical adviser for Id, but it just didn't work out. Probably for the best, as the divided focus was challenging.
"If they don't want me to talk on stage at Quakecon next year, we'll just have to fill up the lobby like the old days. :-)"
Replies
Now when do we get the id tech 5 SDK ?
:poly124:
as long as he still talks at quakecon, him leaving id is OK with me
On the upside, Carmack will make a huge difference at Oculus. Can't wait to see what they come up with now.
I don't wanna spoil it for you, but id Tech 5 wasn't that impressive compared to other tech on the market at that time.
For me it looked like Carmack made one step forward and two steps back compared to what he have done in id tech 4.
I've heard Rage described as a sales pitch for a bad answer to a stupid question.
Megatexturing is a neat idea, and probably an awesome technical problem for somebody who's a dev genius like Carmack or Sweeny to sink their teeth into, but it sounds like it doesn't make for a very workable art production pipeline... at all.
It certainly have it uses, like terrain painting. But idk, if pushing in everything in game was an good idea.
I don't know how assests were made for RAGE, so I won't talk about it.
I was more referring the fact that Doom 3, had dynamic lighting everywhere (at the time where industry was stuck with static lighting). And in time of Rage, industry was trying to move to as real-time rendering as possible, and id tech 5 was step back by using static lighting almost everywhere.
Combine it with MegaTexutring that takes huge resources to develop, and we end up with not so friendly and fast workflow.
I think most of it was done in modo :poly121:
you ask "why"? it's obvious they didn't know how to adapt themselves to these modern days, nor in tech wise, nor in gameplay, nor in art design. they arrogantly thought their "id tech" and the megatextures were great... but in reality it was an illusion, a total fiasco. The best comparison in tech terms: UE3, CryEngine, Frostbite, etc.; and vs gameplay design... err there are too many games to learn of!
"Games are always to innovate or to die".
Their moment of glory passed away a long time ago (Doom 3). I don't see a good future... nor with oculus. Some people called me negative a long time ago, when the "heart of id" left id, and now see... how ironic is the life.
Hows that?
Every product would be more impressive if it was shipped years earlier. Megatextures are more for artists working on the game to make their lives easier iirc. Its also hard to run it on the limited consoles. it looks impressive still on the pc with the latest patch
I'd say it was two or three years too EARLY. The new consoles would have been much kinder to megatexturing's reputation...also, idTech5 is certainly not the only engine capitalizing on the technique--I'd bet there are other studios taking similar approaches, some even doing it on-the-fly instead of baking it down before run time. We'll probably see the term "virtual texturing" thrown around a lot in the next four or five years.
Either way, it's sad to see Carmack move on from directly working on games. He's changed this industry many, many times over.
Peter Molyneux left Lion Head, and that didn't stop him making games. It may not be the same case, but it's another example of a great individual.
The most important difference is that BF and Trials are not using fully unique virtual texturing. But the tech has been in use in other titles as well.
All the new screenspace rendered everything with pbr shaders looks pretty bad compared to Rage in my opinion.
I'm sure Carmack will really be able to stretch his legs without having to support an art pipeline, and I'm pretty excited about VR in general
its allready there... download it from steam...
rage looked great
carmack = great :thumbup:
Thanks oglu, I didn't know
on a supercomputer with 20 titan cards...
I wonder if they will ever finish Doom 4, I was really looking forward to it.
occulus+omni bring it on :poly142:
[ame="
Carmack is a guy that likes to set up new things and invent. The limits of the industry only let him do that once every 5-6-7 years. The PS4 and XB3 just caught up to the stuff they where doing 4-5 years ago.
If Sony and MS focused less on locking in hardware and more on providing a service that runs on a wide range of hardware/devices, they would have a much easier time keeping people like Carmack in the industry.
I should have seamless experience for xbox live, it should be on my tablet my phone, my PC, my laptop, my media box, it should be everywhere and not limited by hardware. That doesn't mean all games will run on all things but what will run on XYZ device, should. That would allow hardware to grow over time and allow services to evolve around what customers want.
Instead they try to predict the future and lock that future down. If you've ever installed an old version of windows and watched the slide shows of what they predicted would be the most amazing features you're going to love, 75% of it is junk no one used. All that does is chase the inventors away.
People like Carmack are like high performance engineers that work on the cutting edge of aerospace or high performance cars.
Sony and MS want people who can make next years Toyota Corolla slightly more/less boxy than this years model. They aren't interested in carbon fiber, they are happy that they just figured out steel is better than iron.
One is to find new and sometimes quite disruptive rendering technology.
He usually gets these ideas by absorbing an incredible amount of information from seriously varied sources; from siggraph research papers to airplane engineering (normal mapping and BSP, respectively).
The other ability is to write a highly efficient implementation of said technology for a video game engine. The best examples are probably the 2D scrolling in Commander Keen and the software renderer for Quake 1. Or I could add the raycasting engine of Wolfenstein where he was using only 1 ray per column of the framebuffer (thus the engine was 200x as fast as one that was casting a ray per pixel).
The common aspect is that both skills were best suited for the time of software rendering, when almost none of the rendering pipeline was set in stone, so radically new approaches were able to allow for drastic advances in visuals.
But nowadays, there's a lot less room left for radical innovations.
Rasterization and texture sampling/filtering are almost completely hardware accelerated.
Renderers are practically standardized around features like deferred rendering, shadow buffers, normal mapping, physically based lighting and shading, and so on.
Innovation is mostly about adding smaller features, or about inventing a new approximation for raytracing a particular aspect (bounce lighting, area lighting etc.)
His last two big ideas in this age were:
- adapting normal mapping based impostors for realtime rendering, which became an industry standard
- implementing virtual texturing, which became a hw feature on at least X1 but possibly PS4 too
There's just not much left in graphics research at this point, as realtime engines are practically feature complete when compared to offline rendering. Yes, the approximations are very rough occasionally - but that could only be helped by LOTS of computing power, and coming up with different shortcuts is probably not exciting enough for him.
The only big step left was virtualizing geometry, which he was looking into; but virtual texturing almost broke id software, so there was no chance to get support from Zenimax to develop a game with the unbelievable production requirements necessary to fully utilize that feature.
Thus it was a completely logical step to abandon realtime rendering related software development. It's mostly diminishing returns from now on anyway, and the game design aspect of id was on the decline in the last decade too.
With that said, I'm still pretty curious about Doom 4, but I don't have my fingers crossed...