Brian Kernighan interviews Ken Thompson at VCF East

This interview was fantastic in person. I haven’t watched the video yet, but they had mics on the participants, so hopefully it’s excellent.


Excellent - thanks! There’s a fair bit of preamble so skip to the 8min mark for the “Fireside chat”

So, indeed, a great talk! I was a bit surprised to hear Thompson say that B came from a shrinking down of an attempt at Fortran until it fit into 4k, followed by an iterative refinement as he added things he wanted - with no mention at all of BCPL. Whereas Ritchie says

B can be thought of as C without types; more accurately, it is BCPL squeezed into 8K bytes of memory and filtered through Thompson’s brain.

But in any case, the story of Unix and of C is always fascinating. And chess software and hardware too! And I didn’t know TMG (the transmogrifier) is the reason why YACC is Yet Another Compiler Compiler:

As an aside, Stephen Johnson named his system YACC because they already had TMG at Bell Labs!

Here’s a version of TMG written in TMG.

In yet another wiki we read:

Ken Thompson used TMG in 1970 on PDP-7 as a tool to offer Fortran, but ended up creating the B programming languagewhich was much influenced by BCPL.[2]

(Another here on B but with no mention of BCPL)

Yes, I was surprised by that story, as well! Possibly there’s a little from column A and a little from column B, and the “filter” of Thompson’s brain was Fortran-oriented. :wink: Presumably Thompson knows, though!

1 Like