Gary Kildall's CP/M : Some early CP/M history
Thomas "Todd" Fischer (originally posted
to the comp.os.cpm newsgroup in 2001)
Fellow CP/M enthusiasts: Much of the early history of
CP/M's roots has been lost over time. For those with an interest, I've
always maintained that IMSAI was the first company to offer CP/M for
8080-based floppy systems. I had the honor of working with Gary Kildall in
the later half of 1976 and early 1977 in implementing various permutations
of rev. 1.2 and 1.3 on the IMSAI FDC and DIO floppy systems. IMSAI's former Chief
Engineer Joe Killian e-mailed some answers to a number of questions I
posed to him, an excerpt follows:
Was IMS the second, or or was it the third company to license CP/M as
an operating system?
Todd, Glenn Ewing bought one of our first 50 or so machines. He was on the
faculty at the Naval Postgraduate school in Monterey at the time. We were
promising an operating system at the time (and later hired Rob Barnaby to
write it). Gary Kildall was also on the faculty at the Naval Postgraduate
school, friend of Glenn's. (Glenn taught antenna field calculations, among
other things.) When we failed to produce an operating system in a timely
manner, Glenn started talking with Gary about CPM, which Gary had written
for Intel under contract. It took several months of twisting Gary's arm to
get Gary to port it to the 8080. The final success came when Glenn talked
Gary into just separating the I/O from the rest of it, with Glenn promising
to re-write the I/O module for the IMSAI 8080 (which he did). So CPM on the
IMSAI was a joint effort between Glenn and Gary. It was only after all this
that we arranged buying it from Gary. I still assume we were the first sale
after his original Intel contract related to this.
This was all possible only after Glenn helped debug the FDC. Peter Olson had
subcontracted to us to design the FDC end of '76, very early '77, following
my general architecture. It basically ran, but had data errors all too
frequently. We had not discovered the errors with our rudimentary testing.
Glenn discovered them with test programs he wrote, and complained. After
some considerable time trying to get Pete to solve the problem, Glenn
offered and we accepted hiring Glenn to debug the FDC. I flew down to
Monterey several times to loan him a scope & other logistics. He identified
the problem and designed a solution. It worked. * That must have been late
'77. Glenn was a quite thorough engineer and very methodical. That was a
very different approach than I'd seen in Pete or other subcontractors we'd
tried to use. We were delighted to hear Glenn express an interest in joining
IMSAI, and he had the job lined up some months before his retirement from
the Navy. He consulted in the meantime.
Glenn later did the VIO. We had started trying to design using the new Intel
video controller, but when we got the first samples they ran only at 6.5V
and had several other faults requiring work-arounds. We were in a hurry as
always. Glenn examined the function and decided we could fit it on a board
using separate parts. It was tight. Glenn did a great engineering job, and
the VIO was cheaper than a board using the big video ASICs for a long time.
Was Seymour Rubenstein the first IMS personage to contact Gary Kildall? How
did Seymour know of Kildall's talents? Clearly Seymour was not the first to
contact Gary, though he did work out the business deal.
Clearly Seymour was not the first to contact Gary, though he did work out
the business deal.
When was Rob Barnaby first brought on board with IMS, and by whom?
I expect mid to late summer '77. Rob was hired by Jan Vath, with some
hesitation since he had been out of programming for a while. We talked about
giving him a try, and if he didn't work out we could "Can his ass", phrase
courtesy of Bruce (Van Natta). It floored both me and Jan
Vath once when we were having lunch with Pete Olson, and there in front of
Jan (my manager at the time) Peter Olson told me I should come and work with
him, he could double my income. Later Pete was to co-found Octel.
Footnote: Joe's salary at the time was about $2500 per month;
trivial given his immense talent and contribution to the founding of IMSAI
*Note 1: We were shipping the IMSAI FDC
systems with CP/M by Spring of 1977, so Joe's memory is challenged with
regard to this event. -trf