In the beginning, 3D graphics software was expensive and was running on expensive SGI machines. It was used by few large (read: rich) companies that could afford it. A whole culture was born, but there were very few people that could afford to learn and use this software…
Then in 1990 Autodesk 3D Studio came along (and Lightwave later jumped from Amiga to PC). That’s when 3D became “affordable”, but not as cheap as MS Word…
When 3D Studio MAX 1.0 came out in 96, it had the same price as today.
Autodesk 3D Studio already had more users than ALL other 3D packages combined, not counting the illegal followers which were probably millions.
So 3D Studio “invented” the mass-market of low-cost 3D applications, especially as games started going 3D in the mid.90s. The Hollywood market is a “niche” – you can sell just as many seats as there is demand for, but you cannot enlarge the market artifically and it is not growing as fast as the game development industry.
That’s why Alias and Avid came out with their “democratic” prices, trying to capture the mass-market (games, visualization, product design, architecture).
They seem to have some good results in games, but I would hate to use Maya on an architectural project for example… Some tools are just not made for certain tasks. IMO, Maya was designed as a TD’s dream application and it does not fit so well in non-VFX environments. If I had to switch applications today, I would go with XSI (but I do not have to, so I am staying with 3ds max of course )
The price of 3ds max is not too high taken into account the development costs and the fact it used to be (and I think still is) the only of the major 3D applications that actually generated revenue throughout its existence.
Maya has NEVER been profitable in the past (before slashing the prices), and I have no idea whether is has become profitable after that (but it might have, if being bought by the owners of Safeway is any indication )
Companies that are professionally using 3ds max can afford the cost as it pays back very quickly. I *suspect* that the wide spread of illegal Max copies had a certain “PLE” effect. I don’t count gMax as a learning edition of Max, of course, and do not advocat piracy, I am just looking at the facts. The 30 days demo is a good step, but I think they could do more…
Linus and OSX are still too small markets, and 3ds max was not designed to run on other OS. In fact, when it was designed to run on Windows NT, NT itself was hardly available, and when it became available, it supported about every CPU out there! (later Microsoft killed support of non-intel platforms because there was no demand to justify the costs of development).
So from the standpoint of 1993-1995 when Max was coded, the reliance on Windows NT was a visionary and pioneering thing!
The reason why the other 3D applications have their Linux or OSX versions is that the pipelines they are used in have traditionally their IRIX roots. With SGI slowly disappearing from the map, porting the pipeline tools to another UNIX-based system is more logical than switching to Windows. Still, a large number of workstations with Maya and XSI are running Windows XP…
Both from technical and business standpoint, it wouldn’t make any sense to port a Max renderer to Linux.
(Autodesk did this with the 3D Studio renderer for SGI in the old days which was never really used and turned out to be much slower on an Indy than on a Pentium 90 anyway).
ALL plugins used by the renderer would have to be available for Linux, meaning that without AfterBurn for Linux for example, you could not render your volumentric out of Max and so on.
NOTE that mental ray stand-alone already runs on Linux (and about any other OS and CPU in existence!), so you can already render out of Max in a mixed network rendering farm. (disregarding the price of render nodes, of course!)
Using an intermediate scene description (like mr does), you could have a stand-alone renderer like VRay.exe or the future Brazil 2 running on another OS (I am not claiming that Chaos Group or Splutterfish are planning Linux versions, I have no idea what their plans are), but you would be only capable of rendering what can be described by their intermediate scene description format… I just don’t see Discreet spending money on porting the Scanline renderer to Linux, and I don’t see anyone supporting a renderer that does not support all 3rd party plugins in existence…