Working Up

by Dwayne Phillips

Some 25 years ago, I learned something that has stayed with me. The reason it has stayed with me is that it still holds true: Don’t depend on live demos.

The title of this post is a statement made often by one of my advisors in graduate school way back at LSU in the mid-1980s. Dr. Charles Harlowe went over this again and again. He was the head of the Remote Sensing and Image Processing (RSIP) laboratory. The RSIP was an inter-disciplinary lab comprising engineers, computer scientists, forestry people, wetlands resources people, and so on.

Way back in the dark ages of 16-bit and 32-bit minicomputers, we took multi-spectral images from the air and tried to have computers make sense of them. We had some advances, and a bunch of us got graduate degrees.

This was one of Dr. Harlowe’s favorite statements. The lab ran on funded research. Everyone few months, a government sponsor visited to learn how we were coming on the research that the sponsor was funding. As graduate students, we wanted to take the sponsor into the computer room and show them our results on the big display. I think the display was a whopping 24-inch monitor connected to a computer that did nothing but drive the display.

Dr. Harlowe wouldn’t allow such. Instead, he made us take photographs of the screen with slide film, create 35mm slides, and prepare presentations for the visitors. We had a special camera and attachment that allowed us to take the photographs. Then we had to take the film to a photo store (they had such in those days). This was all a pain and ate time. So what? Time was the one thing that we didn’t have as grad students.

It turns out that Dr. Harlowe was right. One day a sponsor was to visit at 10 a.m. We had the 35mm slides ready to go. At 7 a.m. the disk drive crashed on the computer and had to be rebuilt – an eight-hour process. While the system administrator, we had those guys back then, rebuilt the disk, we showed our slides to the happy sponsor.

Today? Disk drives don’t crash. Hence, we only have to worry about the 999 other things that could go wrong with a live demonstration. Some of the ones I have seen:

A key person is ill and can’t make it to the demo
Someone deleted a few files the night before just to clean up the disk
The system administer changed a password the night before
A peripheral breaks
Dead batteries

Conclusion? Thirty years later,

Don’t plan on a live demo

Computer hardware is much more reliable today than in the mid-1980s. Things, however, still go wrong. If everything goes right, do a live demo. Nevertheless, have a backup plan – photographs, PowerPoints, videos, etc. Have what will be an appropriate substitute for a live demo.