P.S, Paul. I had no problem negotiating the CalTrain – AmTrak connection. Thanks again for your help.

Many thanks for your extensive reminiscence. One of the joys of working in a young field such as programming languages is having one’s lifetime overlap with those of its great founders; one of the sorrows is watching them pass from among us.

The closest I ever got to meeting John Backus was giving an informal talk to John Williams, Ed Wimmers, and Alex Aiken in what had been, before his then-recent retirement, Backus’s office at IBM Almaden. The subject of the talk was my theoretical Ph.D. research on integrating assignment statements into functional languages. I learned from your reminiscence that this topic would have been of interest to John Backus.

Interesting to see more pieces of the puzzle. The only time I ever spoke to John was a while after the first ACM-sponsored Functional Programming Language Conference. It was to invite him to look at Burge’s Recursive Programming Techniques and other material, because I felt he’s overlooked some useful resources. He obviously had, or at least you had!

So you got your hands on McG! I always thought it was a version of Strachey’s General Purpose Macrogenerator [Computer Journal 8, 3 (1965), 225-241] that Peter Landin and Bill Burge used as a bootstrapping device to spit out assembly-language code and ISWIM SECD “machine” codes for assembly on a conventional machine. I could be off base about that, but it is my recollection from conversations with Burge. Roy Edwards,was the lone Sperry Univac developer on ISWIM after Burge went to TJW and Landin had gone up to MIT for a while (where PAL was done by Art Evans, based partly on Landin’s ideas). I think Roy’s back in the UK but he would remember better than I (the PM for the project when it moved from NYC down to Blue Bell).

Of course, McG was applicative in an interesting formal re-writing/string-computation sort of way. I figured out a version that used single operands and something like reverse-Polish notation but I never built it when I realized that cons cells work better.

Thanks for the extensive reminiscence. I don’t think there was much knowledge of that work or the fundamental papers outside of IBM. It i sgreat to have the progression described and filled-in as you’ve done it.