,
17 tweets,
6 min read
Read on Twitter
Let's take a minute to talk about Atari Pascal — the implantation of the Pascal programming language for the Atari 8-bit computers. 1/x 
I recently received and scanned a bunch of Atari internal documents, meeting notes, and memos regarding the creation of the language. The language was released in 1982, published by Atari Program Exchange, which you might call the "app store" arm of Atari.
It was not published as a slicker, Atari-branded product, like Atari LOGO, BASIC, and Assembler Editor were. Now that I can see the time and financial investment that Atari put into the product, this is pretty strange.
Atari Pascal was developed by MT Microsystems. It's based on MT/PASCAL, a CP/M product from the same company. Wikipedia says that "MT Microsystems wrote Atari Pascal with a planned "super Atari" 8-bit model in mind…with 128K of RAM and a dual-floppy drive" but that is not true—
because we now can see the original functional specification (archive.org/details/AtariP…) which specifies a standard Atari 800 with 48k. The final product DID end up requiring two floppy disk drives, which really did limit the subset of users who could run the software. And yet...
As I said, the language was released in 1982. This is odd because MT started creating the program in late 1979 or early 1980. It is unclear if they stuck to the timeline of finishing the software by October 1980. If they did, what accounted for the gap before the Mar 1982 launch?
Let's talk money. MetaTech preferred royalties, and while Atari seemed to debate the idea internally, the company told MetaTech firmly that it preferred a flat fee. MetaTech asked for $55,000 total over the 8-month development schedule. That's just about $200,000 in 2018 dollars.
It's unclear if Atari agreed to that amount, or negotiated down. Here's MetaTech's entire 15-page development proposal for the language: archive.org/details/AtariP…
If MetaTech had gotten royalties, they wanted $30K upfront then $5-$10 per unit. This would have been a better deal for Atari, as there's no way Pascal sold 2,500 copies as marketed by APX.
MetaTech seemed to be pushing hard to also create a version of FORTRAN for the Atari 8-bits. (archive.org/details/AtariP…) This didn't happen; I don't think Atari was interested. No other company released a version of FORTRAN for those machines either. (atariwiki.org/wiki/Wiki.jsp?…)
Version 0.0.0 of Atari Pascal was delivered to Atari, with 50 pages of documentation, by June 2 1980 (archive.org/details/AtariP…) Making me think they probably hit the October deadline with no problem.
As I said, Atari Pascal was based on Pascal/MT for CP/M, which made it pretty much-sort of-mildly compatible with UCSD Pascal, the standard. A four-page doc gets into the soecific differences between the two. archive.org/details/Compar…
In January, I'll contact the guys who created MT/Atari Pascal for an interview. Several other versions of Pascal also exist for the Atari 8-bits (atariwiki.org/wiki/Wiki.jsp?…) I've interviewed Norm Draper of Draper Pascal and Tom Eckmann of Kyan. (ataripodcast.libsyn.com/size/5/?search…)
Fin.
implementation :)
Update: two people worked on Atari Pascal. Michael G. Lehman did most of the work but died last year. Wink Saville worked on the Pascal linker, but didn't know the answers to any of my questions. Wink started on a KIM-1 and later was a developer of the PCPI Applicard and one of
the principles in Meridian Data where they developed CDROM publishing systems and other projects. So, certainly an interesting guy to talk to, but not with the answers about Atari Pascal and fortran.
