COMPUTER MEMORIES, CHAPTER 75

WORK TOGETHER TO GET THE JOB DONE
At one company where I worked, the programmers would submit the program they were writing, along with test data, to the computer room for a test run. The printed output, along with comments written by the computer operator, told the programmer how the program test ran on the computer, so the programmer could see what his next steps should be.

Too many times the program test was rejected because “this wasn’t spelled right,” “the third box on the left wasn’t checked,” and on and on. As a result, sometimes it took a day or two to get it all filled out right and accepted, so that a few minute program test could be run. In the meantime neither the programmer nor the operator was being utilized to their potential. Obviously, this was long before there were computer terminals for the programmers to use, and computer operating systems that could be instructed to do this and that with a program.

I had taught several programming classes, and years earlier had been an IBM EAM equipment operator, but now was neither a "stored-program" programmer nor a computer operator, but I was asked to look into the problem. I was given a copy of the instruction book, which consisted of maybe a hundred pages of directions on how to submit a program test job, how to fill out the sheet that would instruct the computer operator how to run the job (number of tapes of test data and blank tapes, that were to be mounted), and what was to be reported to the programmer about the results of the test run.

I soon found that the machine room supervisor had his operators trained to follow that instruction book to the letter, and use any excuse they found to insist that the programmer had filled out the control sheet wrong.

Rather than find a way to reject the job, I challenged the operators to find ways to get the job done, including a long forbidden trip to the programmer’s desk, if needed, to ask a question, to make sure what they intended. The operators were thrilled to be able to do that, and they told me they could have run most of the jobs anyway, regardless of what the control sheet said, but their instruction were firm — follow the book or else.

I replaced the inch thick instruction manual with one sheet (actually!) that said something like, “Let’s all work together and see if we can get the job done,” and delivered it to the Department Head. The operators were happy to be able to show their intelligence, and with the help of the computer operators, the programmers learned how to give detailed instructions that told the operator just what he needed to know.

IBM 7080 COMPUTER PROGRAM CONVERSION
In the late 1960s one of the large Aerospace firms in the Los Angeles area had an IBM 7080 computer that was getting older and older, but was still running programs that were very important to the company. To help lengthen its life, they had built a special computer facility, and only IBM Customer Engineers were permitted in that room, in hopes no one would bump into one of the units of the computer, jiggle a wire loose, and cause it to have a problem. Since spare parts were becoming less and less available, when some company back east replaced their 7080, the machine was purchased and moved to Los Angeles, just for the parts.

Finally they determined that since the IBM 360 System (installed in another computer room) was so successful, they would translate, rather than reprogram the 7080 programs to the 360 language. The vastly increased speed of the 360 system, would make up for the inefficiencies of a translated program. The company I was working for had experience in translating programs from one computer system to another (but not 7080 to 360), so we were asked to “look” at this job. The main thing I remember is that they now had tray after tray (tens of thousands) of IBM cards containing programs for the 7080.

I especially remember one very important program that originally consisted of only 200 or 300 cards. The problem was, over the years so many changes, additions, patches, subroutines, and corrections had been added, that now that one program filled two card trays — several thousand cards. Of course there was little or no documentation, and only a few comment cards that told very little about what the program was supposed to do. It became obvious that even if a program translation was successful, the result would be a much less than desirable set of programs.

After our meeting, and a few estimates in the many tens of thousands of dollars to convert the programs, the company decided to bite the bullet, and re-design and re-program the application using well-proven 360 System file and program techniques. My company was happy with that decision, as we did not want to do a very difficult job that was guaranteed to create a product that would be somewhat less than desirable.

THE END, REALLY !
I am not sure if this is the end of Computer History as I lived it, or if it is all of Computer History that comes to mind at this moment. If you wish to confirm, change, question, disagree, discuss any item in this report, send an E-mail to jimhum@sbcglobal.net.