As promised, here's the story of what happened to the source code for the Gates-Allen-Davidoff 4K BASIC interpreter for the MITS Altair.
The three authored the software that provided one of the foundations for the PC business, and we interviewed Davidoff, who worked for 'Micro-Soft' as it then was, here yesterday.
Let's pick up the story in 1987, twelve years after the historic 4K BASIC had been completed. With the Altair long gone, French journalist Andre Warusfel asked Gates if he could have access to the source code. Gates agreed, but the source listings never arrived. Another French journalist, former editor of the now-defunct programming journal EXE and Pataphysics historian David Mery took up the pursuit over the next decade.
"It is this very program which started the micro-computing commercial software industry," notes Mery in his account of the saga, Stalking Bill Gates. "The PC software industry has a short history and we should have access to the milestones of its evolution."
On numerous occasions Gates promised Mery that he'd release the source, without actually doing so. But with the Millenium Bug safely negotiated, things begin to look up. Your humble correspondent has a tiny walk-on part at this stage, tagging a question onto an interview Bill Gates did with a UK trade weekly. This time, Bill sounded much more positive:-
"We were digging around for this recently and found it. We will get the source code up on the Web - anybody who specifically wants a copy can just ask," he pledged.
Well, although the source code isn't on the web, and certainly isn't freely distributable, it is now public. Sorta.
The source of the sauce
The source code to the 4K Altair BASIC resides in the Pusey library at Harvard, although its arrival there seems to have been through chance. An English programmer Ian Griffiths has seen it, taken notes, and has provided some valuable detective work in his account here. It's a great read.
Far from being donated by Microsoft, the source listing in the Pusey library was actually discovered by accident, behind a filing cabinet, by Gates' old tutor Harry Lewis. Lewis is a Dean at the college and Professor in the computer science department. The Pusey listing one of several copies. But, concludes Griffiths, "The implication ... is that Bill's total input to the preservation project was his act of absent-mindedly dropping a listing somewhere in Harvard in 1975, and then granting permission for it to be copied and archived in 1999."
Griffiths was not allowed to make copies himself - which we guess is entirely consistent with Microsoft's 'Look, Don't Touch' policy to its source code. But he did make extensive notes.
And one we were particularly delighted to see. The source contains the following comment:-
00560 PAUL ALLEN WROTE THE NON-RUNTIME STUFF.
00580 BILL GATES WROTE THE RUNTIME STUFF.
00600 MONTE DAVIDOFF WROTE THE MATH PACKAGE
even though Davidoff isn't credited on the archive. Monte doesn't own any football teams, and may not be a billionaire (he may be, we were just too shy to ask), but it's nice that his place in history should now be confirmed.
But soft! The tale gets even more interesting. Griffiths' friend Reuben Harris, a programmer based in London, has been disassembling the earliest Altair BASIC he could find. It's the same 4K interpreter written by Gates, Allen and Davidoff, and the latter was evidently tickled by the project when we spoke to him on Thursday evening.
Reuben tells us the results should be available in the next week or so. Stay tuned... ®