I wrote one story. I plan to write another one. But first I have to insert this one. My first story for HCLE was about me, computers, and college. The story I’m aiming at is about me, computers, and a lot of engineers in a large aerospace corporation. But first I have to write about slide rules.
Few of us were lucky enough to be born into the era that resulted in us being required to learn how to use a slide rule, and then not have to use it – but I used mine anyway.
High school in the mid-seventies was happening as the country asked for math, science, and engineering students. Curricula were based on an earlier decade, but the current events were moon landings. Computers were used for some of the Space Race, but much of it was done on slide rules. The teachers were familiar with them, and computers were so inaccessible that slide rules were the obvious accessory for any geek heading to their college placement classes regardless of whether we’d use computers. I still have my yellow metal Pickett that was so sensitive to heat that the slide would slip out on hot days. No wonder many preferred wood or plastic.
I carried mine off to college, but only used it for the first quarter or the first year. Mainframes were available and the college was encouraging as much programming as possible. The nation needed programmers.
After graduation I was lucky enough to land a job at Boeing, back when Seattle wasn’t trendy. I think they hired me because I didn’t mind rain.
The corporate environment in the early eighties was advanced enough that almost all calculations were performed on a computer or at least a hand calculator. But the mainframes (Harrises and Crays) were only accessible through card decks, or across 300 baud dial-up modems, or by driving to the computer facility with card deck or tapes in hand. There were mini-computers (PDP 11/70) within walking distance, but they were limited in power and storage, tended to crash, and required on-site site coordinators to manage necessarily neophyte users.
Computers were doing the work, but engineers continued to be the source of judgment. Data was printed, but graphs were plotted by hand. Raw data is rarely smooth and curves were drawn to describe trends from which the judgments and decisions were made. A good engineer had a “calibrated eyeball” and could apply experience to winnowing out bad data, and knowing when a difference was significant or not. And that’s where the slide rule ruled even though it wasn’t used. One of the characteristics of a slide rule is that a calculation is put in context. An answer isn’t just a set of numbers. A slide rule can’t be read that closely. What can be read is the answer relative to slight changes in the numbers used. Does it move the slide a lot or a little? Is that within the accuracy of the data that’s being used? Garbage in / garbage out existed before bits were used.
The computer is more accurate, but it provides less context. To provide the same context requires many more computer runs, which take time to setup, conduct, review, summarize, and understand. The experienced engineers lamented those of us making lots of computer runs because we were missing the opportunity to learn critical judgment skills. Having a proper sense of perspective is important when answers can’t wait.
I tried to practice a bit of their technique. It is hard to describe such subtleties but I can point out a simple one. For those who still know how to read a slide rule it is very easy to see that pi squared is about equal to ten, or that the square root of ten is about equal to pi. It isn’t exact, but it gives a good feel for how an equation may be solved. It is quick to notice that the answer is within a couple of percent, which is useful considering that most real world data varies much more than that. Doing the same comparison on a hand calculator isn’t as obvious. Yes, it can be done, and to great accuracy, but there is no visual clue that the relationship exists. It is like browsing through a library versus doing a search online. Context and perspective are lost. The search may be more efficient, but it may not be as effective.
Change happens rapidly and within two years mini-computers like the PDP 11/70 were arising in labs throughout the facility. They were installed but for anything that is as safety conscious as aerospace, any tool must be calibrated and verified. One of my jobs was to verify that the results from one computer closely enough matched the results from another computer. Sometimes the computers were too slow, so while waiting for them to crunch through their next test cases, I’d conduct preliminary comparisons from the earlier results by using my slide rule. The lab technicians responsible for the new equipment and whoever was trying to justify the new equipment were not amused.
We know the computers won, but their entry was delayed because there was no direct replacement for the previous process’ judgment. The new process obviously works well enough, but an argument can be made that the work was more efficient and effective when it relied on human judgment checking a few conditions rather than massive computations covering the entire flight envelope. Some of the reluctance to the new equipment wasn’t because it didn’t work, but because the equipment tended to rely on people who had to learn how to run the equipment first – time that could have been spent gaining a broader experience. And yes, despite my slide rule, I was one of the young guys cranking out thousands of runs.
That background explains a bit of why desktop PCs, and eventually laptops, weren’t as readily welcomed as they could have been. But that’s the next story.