Richard Nelson would be one of the best to talk to about this, as he was in the industry going way back, and was the president also of CHHU (say "chew") the Club of Hewlett-Packard Handheld Users (I hope I got that right) which I only found out about and got interested in joining right about the time it was dissolving in the late 1980's IIRC. It has only been recently that I acquired a bar-code scanner, but I understand bar code was nice for sharing programs at interest-group meetings (as you could just pass a sheet around, or use a photocopier and give everyone their own sheet) and in publications like the CHHU Chronicle. Otherwise programs were usually shared by text printout, not cards or tapes or other media, and each user wanting to use the programs simply keyed them in.
Keep in mind however that memory was very expensive in the 70's, and calculators didn't have much. There were contests to see who could write the shortest solutions to various problems, and I heard stories of situations where just when you think you have the absolute shortest possible program, along came someone with another entry that was only half as long, and by that time you couldn't even figure out why it worked at all. There may have been slight exaggeration in these second-hand stories; but the point was that for the kinds of things calculators were used to solve, you could usually add more time if needed, but the memory limitations were unyielding. With such small memory amounts, it didn't take more than a few minutes to key in a program if you lacked a faster way to enter it.
Also, people (including myself) usually got programmable calculators to do their own custom applications, and it was not expected that much canned software would be available. Even home computers at that time were bought with the expectation that the user would do much of his own programming, whereas today the common computer user doesn't know how to do any programming at all, and PC applications are the result of thousands of hours of work by teams of professional programmers.
HP calculators were indeed more expensive than other brands, and that was what initially kept me from buying an HP, even though after all my research I definitely would have gotten one if money were no object, because it was clearly better in many ways. Even HP's keyboards felt
like quality, and were reliable, unlike TI's. RPN made better use of memory too, and I liked the fact that the parentheses were gone. After a few years of using the TI-58c and 59, I got an HP-41cx because the TI's had no capability to interface to lab instruments, whereas the HP did it with ease, in fact, much more easily than a PC. At first, I could have gone either way, AOS or RPN, but the instrumentation and control area quickly made RPN come out the clear winner. TI always liked to claim that the algebraic system was better because you could enter an equation just the way it appears on paper; but the truth is that in real-life (ie, not school) applications, you usually don't have
an equation in front of you. Repeating a post of mine on the HP Museum forum:
In the past I worked at a local college as a math tutor and looking back at that experience I wonder why educators do not more readily embrace RPN. Instead of teaching students calculator syntax ("you missed a parenthesis") to make the equation look like the one in the book, that time could be spent reinforcing the order of operations and actually teaching the students to become more self-reliant in doing the math. Seems like everybody wins.
And that's part of engineering. Teaching them to blindly throw the numbers in the chute and turn the crank, ie, plug numbers into a pre-written equation, is not
the way to make them understand, and it does not
produce real engineers. What happens when they lose or forget the equation, or have an application for which there is no pre-written solution? They're lost. My 41 contains programs I wrote for the things I do frequently; but the majority of calculations I need to do come up without warning, and there is no equation in front of me
. This is real life. Thinking about the problem, I figure out the steps as I go, whether it's to take A, raise it to the power of B, subtract C, divide the whole thing by D, add E... whatever-- you get the idea. I've hired a lot of technicians and a few engineers, and I give them all a little test, and I have to say I have never found an applicant who learned to do even basic circuit design in school and calculate circuit values. Somewhere--I wish I could find it now--I have an editorial I cut out from one of the industry magazines about this, saying that the industry is telling academia, "You are not turning out the kind of engineers we need. We need them to be able to do this, this, and this..." and academia responds, "Look, you know your field, and we know education. Leave the education to us, 'kay?", so the problem persists.
The TI-59 came out in 1977 with its single port for dropping in a software module. It only held one at a time, and the software in the modules was all written in the user language. The HP-41 came out in 1979 and offered four ports for modules, and more flexibility. Modules could have code written in the user language or in assembly which dramatically improved performance and gave more capability, and there were more types of modules possible-- RAM, ROM, timer, I/O, etc.. A module could even wake up the calculator to do jobs when needed. There were dozens of these, and yes, they were trustworthy. Today new modules are still being produced for the HP-41, both hardware and software, going much further than the technology allowed in the 1970's and 80's.
The users' groups added a lot to the value of the calculators, especially the HP-41. I doubt there was much financial incentive to contribute to the program libraries. People would write programs to do something they needed, felt good about it, and wanted to document and share the program.
I did not get into the programmable calculator scene until Dec 1981.