Computer Oral History Collection, 1969-1973, 1977
4
Grace Murray Hopper Interview, July 5, 1972, Archives Center, National Museum of American History
( ?). Plus the fact, since that was built in it, it had to have complete
generality ( ?), but it also had to meet the full generality of the machine, so
( ?) twenty three digits. You could, I only needed five, but there was no way of
telling it to stop short of flat digits, it always gave full generality. As a matter of fact, I
pulled the relays. I pulled the lower twelve position relays in the unit, of the constants. So
it would only give eleven digits. (LAUGHTER). It got left out one day when somebody
was doing something else and Aiken decided we couldn't pull relays any more. The
reason those units were pulled out was that they always delivered complete generality
and no individual problem ever required complete generality. So what happened was, we
began to develop routines in our note books. And I had, let's say, a cosign routine for
positive angles delivering ten decimal places. Bob Campbell might have one for angles
less than (pi over two?) delivering eight decimal places and we'd yell at each other, can I
have (your essential?). We had relative coding at that time. I think I turned over some of
those sheets to what's her name which are relative coding essentially. And we copied
each others routines which was one of things that was the basis for the concept of the
(compilers?) you see. But those units were, we stopped using long before they were torn
out of the machine. And they are the wrong kind of a thing to build into a machine. Now
you can build in something like an index register. That's fine, for that's control of the
program, not the numerical, but you build in something that's numeric and you're always
restricted with complete generality but you never require that for an individual problem.
Because that's what a problem is, it's something less than complete generality. You
specify this…(voice fades out).
TROPP:
Of course that brings up I think, a philosophy that pervaded, at least as I see it, some of
the early hardware machines in terms of Aiken's ideas about what the machine was going
to be and at some point in time that became fixed and the machine was built. And even if
you saw later that some of these things ought to be changed, it was too late to do it.
HOPPER:
I don't think it went, I think it went too fast for that. Those things were built under a very
heavy wartime pressure.
TROPP:
I guess that's one of the problems, getting back to the, trying to recreate the atmosphere
of the War.
HOPPER:
Don't forget that Mark II was built under wartime pressure. Get that thing built fast. Use
existing components. Nothing new. Get it built. Get it running. It was critical. That was
done awfully fast. The whole design and development.
For additional information, contact the Archives Center at 202.633.3270 or archivescenter@si.edu
Computer Oral History Collection, 1969-1973, 1977
5
Grace Murray Hopper Interview, July 5, 1972, Archives Center, National Museum of American History
TROPP:
I remember discussing with, I think it was Dick Bloch, the problem of getting the relay
built and designed for that particular machine.
HOPPER:
You will find that Dick Bloch was the mathematician, he was the theory man whom you
get much more of the practical building and everything. The realist was Bob Campbell.
He was the one who really built it.
Bob was both theoretical and practical. He could design an adder and build it. But Dick
used to change the instructions on the Mark I over night. If he thought up a nice
instruction to make one of his problems easier, he'd put it in and none of our tapes would
run the next morning. (LAUGHTER).
TROPP:
I'm sorry Beth, I'm getting away from your set of questions.
LUEBBERT:
I'm fine. I just wondered…
HOPPER:
I told her I was going to write a book. (LAUGHTER).
LUEBBERT:
Going back to Babbage for just a moment. Where did you find out about Babbage?
HOPPER:
Aiken gave me his book to read. He by then had a copy of "Lives of a Philosopher".
LUEBBERT:
And that's where you…
HOPPER:
Which he had written. I had never heard of him before that.
LUEBBERT:
For additional information, contact the Archives Center at 202.633.3270 or archivescenter@si.edu
Computer Oral History Collection, 1969-1973, 1977
6
Grace Murray Hopper Interview, July 5, 1972, Archives Center, National Museum of American History
And you went further then Aiken?
HOPPER:
?
LUEBBERT:
You went further in understanding?
HOPPER:
Well when I wrote the first chapter in this, I started to find out what had been in existence
earlier. You notice I find some of the American machines and some were Babbage
machines. I didn't have time enough to go as far as I wanted to, but I did round up and
read most of that material. A good deal more than I wrote in my first chapter.
TROPP:
Had you spent much time looking at some of (Hollerith's?) developments?
HOPPER:
No I didn't. There was nothing available in the Harvard Library and I didn't travel in
wartime.
You've got to remember all that was done in wartime.
TROPP:
That's right.
HOPPER:
We were limited by wartime restrictions.
LUEBBERT:
I was interested in how you got to Harvard in the first place?
HOPPER:
Ordered by the Navy.
LUEBBERT:
For additional information, contact the Archives Center at 202.633.3270 or archivescenter@si.edu