In 1969 there appeared at General Electric one Richard S. Bloch, whom I
had known from earlier days. He intended to design a new line of
IBM-compatible computers for GE to market. I was assigned to his team,
despite an overwhelming aversion to such an approach, due to:
- RCA was unable to make (essentially) the same scheme work, and I
knew the inside story of that, having been invited twice by RCA to be
their head of programming around the time of that effort. I declined
both times, even though the New York recruiter served a lovely lunch of
broiled shrimp and wine in his office. My wife had a hand in the
decision; one of her best friends was dating Ed Donegan, who ran the RCA
effort, so I had some inklings of their troubles.
- I had been a strong influence, while at Univac, in upgrading the
1107 to the 1108 by putting a faster engine in a good software body.
This was successful. The program was approved on an estimate of 30
sold, and they sold over 300.
- From 360 days, I knew that 32 bits as a floating point vehicle gave
much inaccuracy, forcing a larger proportion of double-precision
calculation, with much lowering of speed.
- I had a strong dislike of the EBCDIC code, having worked quite hard
at IBM to replace it by ASCII (and I would have succeeded were it not
for one Vin Learson).
Designing the content of the ACM 70 conference proved a fine way
to avoid being stigmatized by actually working on the project,
particularly during the "Shangri La" design concentration camp in
Florida, where I only had to spend two or three days.
In late 1969 I made a special trip to Schenectady, ostensibly to brief
GE management on the purposes and content of ACM 70. But there I took
the opportunity to defend a number of memos I had circulated to high GE
management, in which I took strong issue with the prevailing plan.
My audience was primarily a Dr. Jack Music, famed as the "nose" for
Hilliard Paige, one of the three high-level GE executives deciding on
the viability of Bloch's plan. Music added me to the investigative
team. As far as I know until now, I was the only member of that team
actually working for the GE Computer Department.
Another team member was Mike Kami, formerly head of long-range planning
for IBM. I was enjoined by Music to not tell my manager where I traveled
or what I was doing, causing our expense account approval lady to
scold me for turning them in so late.
But the reason for this short memoir is to show how very complex the
computer business is, and to restate my theme that it is unlike most
others, in that to be successful it must be run by a management team
that is fully experienced in that field. I.e., contrary to GE's theme
that "A manager is a manager is a manager ..."
So on 1969 Oct 14 I found myself, as Mgr. of Systems and Software
Engineering Integration for the GE computer business, sitting in a
Bridgeport, Connecticut, conference room with the chief planners of this
enterprise.
On my left was the man (name mercifully forgotten) who was to head the
field support group for software. During a lull, this ensued:
- Bemer: "How many programmers do you plan to start with?"
- He: "Programmers?"
- Bemer: "Yes, how many?"
- He: "How many do you think I need to begin?" (an obvious parry)
- Bemer: "About 50"
- He: "50?" (in a gasping voice)
- Bemer: "Have you ever been to Mahwah?"
- He: "What's Mahwah?" (very puzzled)
- Bemer: "It's a town in New Jersey, where IBM does what you are going to
be doing for GE. They have an area of one and a half football fields
full of computers there, entirely surrounded by programmers."
For some reason he avoided me after that. And (strangely) GE sold its
computer operations to Honeywell that next June.
Back to History Index
Back to Home Page