The premise that anyone is going to rent an expensive 3D CAD software package for a single job because a client requires it ignores the often steep learning curve for most programs of this type. Designers typically work on a fixed cost basis or budget. You can't make any profit if you are not productive and you can't be productive using new software for which you have little or no training.
The ASP community seems to be using the business model of an exotic car rental agency. Their pitch is: "Want to drive an expensive foreign car, but can't afford one? Fine, we'll rent you one!" For automobiles this model works, because most of us know how to drive and can learn the special attributes of a particular vehicle in a matter of minutes. But this business model does not apply to a complex 3D CAD software package (unless you are simply adding another seat on a temporary basis). I have "test-driven" enough demo software to know this is true.
I was taught that training costs are a major portion of the "Total Cost of Ownership" equation. It's true that CAD packages are becoming more user friendly, but it can take a year or more (with training classes and daily practice) to become truly productive. As the capabilities of 3D CAD software packages improve, the learning curve (and cost) increases as well.
Personally, I used CadKey for over 10 years, and have used Sol id Works for the past year. I am still working on improving my productivity and see no quick solutions to mastering any new 3D CAD software package (or even a new release of a current package).
If a 3D CAD software vendor is really looking for the next big thing, I suggest it include free extended training class offers with its software. This should also apply to the free upgrades for which we all pay exorbitant maintenance fees.
It would be interesting to see if other readers share my opinions.
Ronald L. Watkins
Our author does discuss training in her article, though in most cases she includes it as part of "support." You are right that training is an absolutely critical component of the ASP business model. It will be interesting to see how well it ends up being addressed over the next several months.
In your magazine, I have seen reviews of Intel Pentium-based systems, but never of AMD Athlon-based machines. In articles in other magazines, I have read that AMD's Athlon processor outperforms the Pentium III in 3D graphics and animation. Why hasn't there been a review or comparison of the two processors in your magazine?
I recently purchased an Athlon-based computer, on which I'm currently running Discreet's 3D Studio Max R3. I'm happy with the performance so far. Why haven't I seen companies that produce editing and graphics workstations switch over to the Athlon?
Huntington Beach, CA
As for the reviews in this magazine, your point is well-taken. We'll certainly keep an eye out for a review-worthy AMD-based machine.
With regard to your more general questions, we asked analyst Jon Peddie, principal of Jon Peddie Associates in Mill Valley, California, and frequent contributor to Computer Graphics World, for his take on the AMD processor. Here's what he had to say:
From a workstation perspective, the AMD processors have the horsepower, but not the support I/O. The reasons are AMD's lack of support (via the chipset) of super high-speed memory (DDR or RDRAM). An even more important reason is that AMD does not have an AGP license from Intel. All the high-performance graphics accelerators are using AGP 4X now to gain access to main memory.
AMD does have a powerful chipset controller, which gives it a good start. However, the AMD processors and associated core-logic shipping today can't support multiprocessor operations, another feature required by many high-end professional workstation applications.
So it's not the processor alone that is the qualifier for workstation status. I believe AMD understands this and is working toward entering the workstation market to exploit the speed advantages of its processors.
That aside, your satisfaction with your Athlon-based machine is a prime example of why much of the benchmark hype should be ignored (given the confused state of benchmarks, the misuse of them by vendors, and the ignorance of what they test on the part of many publications). In this case, YOU are the benchmark. The machine is working for you, and it's cost-effective. You are to be thanked for bringing this issue out into the open.
We welcome any insights you have to offer that would further our readers' understanding of topics discussed in this issue, or that concern the computer graphics industry in general. We may edit your comments to conform to our style and space requirements.
Please address letters to:
Computer Graphics World
98 Spit Brook Road
Nashua, NH 03062-5737