Understanding Computers
Eric Plow recalls what it was like to operate computers at his job at the UNC-Chapel Hill School of Dentistry in the 1970s. He explains the capabilities of the computer and the cost associated with completing a "job," or project.
Interview on 2012-04-19 00:00:00 -0400
Transcript
00:00:00.000
Well a computer was huge, it generated tons of heat.
00:00:04.370
A computer probably would fit in this whole room
00:00:08.475
and it would cost a million dollars.
00:00:11.719
Back then they were just amazing what they could do
00:00:17.410
but compared to what computers can do nowadays they were pretty weak,
00:00:22.801
but we didn't know that back then
00:00:25.651
so we enjoyed the computer power that we had.
00:00:29.065
But because so many people wanted to use the computer
00:00:32.327
the computer had time sharing,
00:00:35.397
so you would punch your program on a bunch of computer cards
00:00:40.666
and some of the programs would be two thousand, four thousand cards long
00:00:47.135
and you'd sit there with a card punch and punch this thing in
00:00:50.409
and if you made a mistake you'd have to throw the card away and re-punch it,
00:00:53.753
so it was very laborious.
00:00:57.582
Then whenever you wanted to run a task, or we called it a job, a computer job,
00:01:04.016
an activity, like if you wanted to compute a statistical analysis of some data
00:01:10.874
you would load the program, you would load the data on cards,
00:01:16.185
you'd bring it to the computer center, this big monster computer,
00:01:20.660
they would run the cards through a card reader
00:01:24.723
and then that job that you submitted would sit there
00:01:29.209
and it could run in five minutes or it could run in five hours or it could run ten hours from now.
00:01:36.345
You had no idea when the job would run. Once it ran you would get output from the job, printed paper.
00:01:43.758
It wouldn't display on a screen.
00:01:47.296
So you would have to sit there and-. Well, you wouldn't wait because you could be waiting there forever.
00:01:53.358
What you would usually do is go study or do something else and then come back in a few hours
00:01:57.791
and hope that your output was there.
00:02:00.234
If it wasn't you'd have to come back.
00:02:03.027
Well the frustrating thing is if you had any kind of syntactical error in your computer program, if you missed a comma or misspelled a word,
00:02:09.499
it would just reject it and give you an error message and you'd have to start all over again.
00:02:15.748
So to run a particular task could require several iterations
00:02:21.826
One of the big things that affected the cost of the job,
00:02:28.289
besides the size of the job and how much computing power you needed,
00:02:31.798
was the priority which you wanted to run,
00:02:36.181
but every time you upped the priority the cost would double
00:02:39.532
and computing back then was not cheap. If you did a simple computer job it could easily cost twenty dollars,
00:02:45.580
which even back then was a lot of money.
00:02:47.759
If you run it several times you're up to a hundred, a hundred and twenty, a hundred and fifty dollars just for one run.
00:02:54.080
So if you ran it priority one you'd be lucky to get the thing back in a few hours.
00:03:00.409
If you ran priority two it would come back maybe in an hour or two
00:03:04.783
but it would cost twice as much.
00:03:07.223
If you ran priority three it would really be expensive
00:03:10.099
but you'd get it much-well nobody had the budget to run priority three.
00:03:13.921
We usually ran priority one but our jobs at the dental school were so large and so expensive
00:03:22.014
and so time consuming that we ended up running a lot of stuff priority zero
00:03:27.178
which meant that it would only run at night.
00:03:29.751
That's the only time they would run it.
00:03:32.468
So I spent many, many nights
00:03:35.639
getting there at midnight, and usually I'd be one of three or four people at the computer center,
00:03:41.777
and back then you could sit your job in and it would come back within a few minutes
00:03:48.034
and you could correct it, put it in again, and just get a really good turnaround,
00:03:52.408
which you could not do during the day. So you ended up working a lot at night and lost a lot of sleep.
This video is an excerpt from a longer interview. Contact the Special Collections Research Center to request the transcript of the full interview.