Basic Philosophy Part II: Computers and Technology

I have been involved with computers for a very long time, starting out with FORTRAN and progressing to BASIC, assembler, Pascal, Ada, C, C++, etc. I have seen companies come and go and seen first hand the mistakes they made. Some well known and very brilliant people made incredibly stupid judgments. Of course, it is much easier to realize how stupid they were in hindsight, but sometimes it was apparent to others even at the time. I have evolved a few simple heuristics or rules. These are all common sense and I am sure you have heard them before. But what amazes me is that people continue to ignore them.

Develop for the Future

The first rule is that any long term project should be developed for the environment that will exist when it is deployed rather than when it is started. So you need to understand and be able to forecast the technological and social environment in which your systems will be used. There is no point in working on solving a problem that will take three years if the problem or your target systems will be gone in three years.

Several years ago I had a friend who was working for a very large computer company (CDC - no longer around). She was part of a large group developing a business system on 16-bit minicomputers. Rather than being impressed with the room full of computers, my first question was why were they developing the system on machines that would be obsolete by the time they finished. The answer was “We got a really good deal on the computers”. A year later the company pulled the plug on the project when they finally realized that no one was going to buy it because no one was buying 16 bit minicomputers any more. So, they lost about 200 person years of time plus the cost of all the facilities, because they “got a really good deal”. No wonder the company went out of business.

Another example is in computer graphics hardware technology. People used to worry about quantizing 24 bit images so they could be displayed on 8-bit graphics hardware. That problem has disappeared since everyone has 24-bit hardware now. But the human eye is more sensitive than the 24 bit displays we currently have so people are looking at ways to create and display High Dynamic Range (HDR) images on 24 bit systems. But it is irrelevant to spend 3 - 5 years to find a software workaround for this, because by then we will have HDR displays. I first saw an HDR display in the Emerging Technology area at SIGGRAPH 2002. There is another, more advanced HDR display system being shown at SIGGRAPH 2004 (http://www.siggraph.org/s2004/conference/etech/high.php?=conference).

So always be thinking about the future.