ALAN: Alan Hoffman
IRV: Irvin Lustig
IRV
You were involved in the early symposia on mathematical programming. What was your involvement?
ALAN
I organized one of the earlier linear programming conferences, not the one known as zero, or the first, but the second. It was in the early 1950s and at that time I'd been in charge of computing work we were doing with the simplex code written under Alex Orden's direction. So, when I organized this conference, I felt obliged to speak myself and I also felt obliged to speak about how wise I was concerning techniques for computing the answers to linear programs. I wrote a paper and it's not my favorite. I've since looked at it. All the predictions I made about the future were wrong. All the comments I made about the present were naïve.
IRV
But a later paper on computational experience in solving linear programming problems made an impact, didn't it?
ALAN
There were an enormous number of requests for that paper and that was before we had copy machines! I couldn't understand it. Harold Kuhn said that it wasn't that people were so interested in what our results were, it's just that they wanted to have answers to problems to make sure that their codes were correct. But the paper has some celebrity with the Mathematical Programming Society. They established criteria for what articles on computation should contain, the style and so forth, and they always cite this paper as an example. The reason why we gave so much information is because we had accumulated so little that we felt obliged to say everything we knew. Nevertheless, I think we designed some reasonable experiments and I think we said what we knew with some grace.
|