For the purposes of this debate, a PC will be defined as a machine sold through retail channels, while a workstation will be purchased through the manufacturer. By this metric, virtually all Intel and Apple machines are PC's, and machines with non-Intel and non-Apple- PowerPC processors are workstations.
Group members are:
Clearly since consumer PC's and game platforms have 3d hardware acceleration today, for the forseeable future they will continue to do so. The question is whether or not there will be a need for workstation accelerated graphics, or will consumer machines provide enough functionality and performance to make workstation accellerated graphics economically unviable.
I would argue that workstation graphics will continue to be needed for the forseeable future, since consumer graphics will not provide suffecient performance across the wide functionality range (eg. OpenGL) required by computer graphics professionals.
To backup up this statement I offer up the following:
1) The consumer market, by definition sells to large numbers of people in order to take advantage of economies of scale. By nature the consumer market will not purchase value added products, such as a graphics accelerator board or game console, for much more than $300. The general market also uses 3d graphics for one purpose only: entertainment in the form of 3d games. Given these constraints, firms must choose to improve game performance as much as possible within the given parts budget. Since games use only a limited set of graphics techniques, consumer firms will not implement a full OpenGL style hardware board with high performance for all functions.
2) In order to accelerate certain graphics features, a large amount of memory, or parallel hardware is required. Even though consumer graphics firms can get the absolute lowest cost per part due to their high volumes, volume discounts only allow the part cost to sink to the price of its materials. Since there is a fixed budget that can be spent on parts, they must limit themselves to only accelerating essential features using techniques that don't require a lot of hardware, and therefore will not be able to accelerate the full range of OpenGL like features. Workstation companies are not limited in this way since they sell at a higher price point, so they can afford to "throw hardware" at the acceleration of certain features.
Based on these two points, I contend that workstation graphics will be an important market for the near future.
I plan to describe the next generation PC hardware, focusing on the Intel - Lockheed-Martin partnership, its technical details, why MMX is so cool, and why PC's will be the low-cost graphics platform of the future.
As I discuss each of these issues, I will focus on the higher-end applications (VR, realistic image/video/movie generation, visualization) and features (improved lighting? very complex models/geometry?). My intent was to advocate a single, high-end, high-cost machine or system as the provider of these advanced features, but I might also explore the alternatives (many lower-end machines, or perhaps a not-so-high-cost machine which is missing some of the very-high-end features), perhaps discussing the pros and cons of each - this might stir up more discussion than if I advocated just the high-end of the high-end. Also, I do think that different future applications require different types of architectures, so this would allow me to get around advocating high-end machine for an application that they are not best suited for.
I'll argue that Direct3D will become the PC standard and that OpenGL will be relegated to niche markets.
Other suggested topics (so far) are below.