The Medley Interlisp ProjectIn the late 1970s Interlisp was ported to DEC VAX computers under Berkeley Unix. These papers reported on the project and its challenges.
https://softwarepreservation.computerhistory.org/LISP/interlisp/Interlisp-VAX_A_Report.pdf
David, a Bostonian in Tokyo.Interesting stuff.
It was painful moving from the PDP-6/PDP-10 instruction set (two 18-bit pointers per word) to the 32-bit computers (one 32-bit pointer per word), since data became twice as large and accesses became twice as slow. But the writing was already on the wall by 1976 or so, when MACSYMA was feeling constrained by the 18-bit addresses. My memory has it that the Multics implementation of MACSYMA was four times larger and ten times slower than the KL-10 one.
Ouch.
(Continued ...)
For reasons, I disappeared from the Lisp universe from '76 until '82. When I reappeared, Yale (Jonathan Reese) had implemented Scheme (called "T"). It was finally usable when we got 68020 workstations.
It was an irritation that despite my diversion to hyperspace, it wasn't until 1984 or so that I had as much Lisp compute available as I had had in 1976.
Whatever. Hats of to the heroic efforts of the folks at the time to implement Lisp on difficult hardware.