Wednesday, January 11, 2023

The mistake-filled demonstration - that worked


When I first started working for the company that I retired from recently, we didn't have a sales department or a marketing department (this was back in the late 1980s). We did some advertising in magazines that were read by people who would probably buy our product but the biggest reason people bought our products was that other people told them about our products. I could spend a long time trying to explain what our product did but I will try to keep it simple.

Our product was a small (less than 3 inches by 5 inches and some even smaller) computer board that could run from batteries and could be used to collect sensor data (temperature, humidity, air pressure, movement, sound or any sensor or transducer that produces a voltage or digital pulses) or to control other electrical equipment (turn motors on and off, turn lights on and off, send commands or data to another computer or the Internet). They were used by many companies and universities but one of our biggest customers at that time was close by - the Woods Hole Oceanographic Institution (WHOI) on Cape Cod. Many times, their experiments were run in remote locations on the ocean, under the ocean or in inhospitable places like a jungle, a desert or the arctic. WHOI's equipment would use our board to run the experiment. It would time the actions of the experiment and collect data from the experiment. Then, they would come back and pick up the equipment and read the data in their lab. Or the data may have been transmitted back to their lab by radio or satellite. Our boards were very flexible because you wrote a small program to organize what what happening when.

So, once in a while, all the engineers (all four of us) would take our latest products to demonstrate their capabilities to our old customers and some possible new customers. These folks were mostly scientists. They would later design elaborate controllers for their experiments but in the early stages of a project, they could buy our general purpose controllers to run their experiments until they nailed down what they really needed. They liked our boards because they were easy to use and very (very) low power so batteries would last a very long time. So, usually, our boss (the owner of the company and chief designer of the boards) would do a demonstration to show them how things worked. He was brilliant but didn't use the boards all the time. But he was a respected scientist himself and knew many of the WHOI scientists so it seemed logical that he would do the demo.

Our boards came with a built-in operating system and could be programmed in BASIC `(and later C and even later a Java-like language) so it was easy to pick up. "Doing the demo" consisted of typing a short program on a desk-top computer and sending it to our board which would then respond by sending information back to the desk-top computer with the information you requested.

He'd start out simply by having it display the current temperature. Then he'd use a little math to convert the default Celsius temperature into Fahrenheit. Then he put his finger on the temperature sensor and show that the temperature had, indeed, changed. Then he'd get fancier. He'd program a simple loop to read and display temperatures over and over as he held and released the temperature sensor showing the temperature change. Then he'd add a delay in the loop so the temperatures weren't displayed too quickly because, really, the temperature doesn't change that fast and if you are storing the data, you don't want to use up all of the memory so fast. With each step, he'd add functionality to make this simple board act more and more like a well designed experimental controller. But as the additional program code became more intricate, he'd start making mistakes. The first time I saw this, I gasped - thinking that he'd ruined the whole thing.

But he'd correct himself right away and quickly make the change he needed. Sometimes, the errors were more subtle and it would take a little longer to fix the problem and sometime the Woods Hole scientists themselves would offer suggestions. He'd always find the problem and, once it had been identified, fix it quickly. You could see the scientists leaning forward when a problem was found and then relaxing back in their chairs as it was fixed. By the end of the demo, they had seen a pretty complete data logging system designed right before their eyes and we usually had a lot of excited people gathered around to order some of our computer boards. They had seen how easy it was to come up with an idea, try and modify it simply and easily.

It was almost like listening to a sermon when the preacher lists a problem or asks a question that the congregation has been wondering about and then shows how the Bible has the answer or Jesus had a parable about just the problem they were wondering about. Some preachers raise their voice and sometimes they would lower their voice and the people would lean forward to hear the answer. But just listing a bunch of answers doesn't make the listener learn anything. Keeping the audience engaged helps them learn how to think about the problem instead of just hearing the answer.

By the way, our boards were available long before the current Arduino and Raspberry Pi boards were available but our boards were a lot more expensive than those new boards. We gave lifetime, free support for our boards and no one made (or makes) a lower power solution than we did. But the company's focus changed and our boards are no longer produced. But I'll never forget the exciting demos our boss did. He knew how to hook a crowd.



No comments: