2 May
2024
2 May
'24
8 p.m.
On May 2, 2024, at 3:50 PM, Lee Courtney via cctalk
<cctalk(a)classiccmp.org> wrote:
The first "professional software" I wrote (almost) out of University in
1979 was a package to emulate the mainframe APL\Plus file primitives on a
CP/M APL variant. Used to facilitate porting of mainframe APL applications
to microcomputers.
I'm still an APL adherent since the late 1960s, but it was probably too
heavy-weight, with obstacles noted elsewhere (character-set, radical
programming paradigm), to be successful in the early days of
microcomputing. Although the MCM-70 was an amazing feat of technology.
Too bad because the language itself lends itself to learning by anyone with
an understanding of high school algebra.
The one professional application APL I heard of was in a talk by Ron Rivest, at DEC around
1982 or so. He described a custom chip he had built, a bignum ALU (512 bits) to do RSA
acceleration. The chip included a chunk of microcode, and he mentioned that the microcode
store layout was done by an APL program about 500 lines long. That raised some
eyebrows...
Unless I lost it I still have the article somewhere: it's the cover story on the
inaugural issue of "Lambda" which later became "VLSI Design", a
technical journal about chip design.
My own exposure to APL started around 1998, when I decoded to try to use it for writing
cryptanalysis software. That was for a course in cryptanalysis taught by Alex Biryukov at
Technion and offered to remote students. The particular exercise was solving an ADVFX
cipher (see "The Code Breakers", the unabridged hardcover, not the useless
paperback). It worked too, and it took less than 100 lines.
paul