System's Programmer (1970s)
This was back in the days when the source for the system came on the tape
and system programmers provided their own fixes. System paging too much?
- rewrite the paging algorithm! A great deal of the advanced functionality
in the VM products came first from IBM's system's programmers. Ahh,
going into work at 3 or 4 in the morning - it brings tears to my eyes.
See the original IBM presentation materials for VM/370 here.
Boeblingen Development Lab (1973-1975)
Very early in my IBM career, I was offered a chance to work at the development
lab near Stuttgart, Germany, and they paid me to go there! It only took
3-4 microseconds to tell them I'd go. I worked in a VM tools group supporting
DOS/VSE VSAM development. Before I left, we had written special simulation
code for a new line of IBM machines - the 43xx machines. Pretty
nifty at the time!
Software Development Tools (1978-1981)
After returning from Germany we decided it was time to move closer to family
in the Midwest - so I transfered to IBM at Rochester,
Minnesota where I worked on MVS/TSO systems to develop a wide range
of tools for the developers of the brand new S/38.
After several years on MVS, I led the technical effort which converted
the entire development environment to VM. This was a larger social
change than a technical change. I learned that it's pretty tough to convince
5-600 people to change how they work - regardless of the merits of your
SNA on CMS (1981-1983)
Again, the lure of Europe beckoned in the form of an opportunity to work
for IBM U.K. at its Hursley development
lab. There, a small team worked on SNATAM - a project to create a fully
functional SNA subsystem for CMS. SNA is defined by a language called FAPL
and our project was to "compile" this code into a fully functional SNA
My role was to write a CMS multi-tasking subsystem and debugger for
the project. I'm sorry to report the CMS product never did get mutli-tasking
support from this.
One of the more notable "accomplishments" at the lab had nothing to
do with SNA. My wife and I noticed early-on that we didn't speak the language
as well as we expected. This led to a great deal of fun as we collected
and wrote the "English" to "American" Dictionary.
IBM BackBone (1983-1987)
After returning to Rochester from England, I joined a project whose goal
was to greatly improve the reliability and speed of the VNET (this
is the IBM internal network - used to transfer gigabytes of data daily
between hundreds of machines world-wide). Working with Pete Hilton (then
with IBM U.K.) we designed and implemented a new version
used on the ten "backbone" machines scattered around the world. Much of
this work eventually was put into the RSCS product. I received an Outstanding
Contribution Award for my efforts.
Advanced Technology Group (1988-1992)
At this point I joined a project consisting of 20 software programmers
and 20 hardware engineers whose mission was to "improve productivity 10-100
fold". This was a very stimulating group - the mixture of software and
hardware people proved a potent combination. The group was responsible
for many innovations and changes at Rochester - both in the AS/400 product
as well as the working environment.
My group chose to try to improve programmer productivity. The thrust
was to convince Rochester development to use Object Oriented Programming
combined with a powerful workstation running Unix. This process took years
to invent, then implement and convince Rochester management to pour millions
into the change. Once again, I found myself in the core group trying to
convince hundreds of developers to move from their comfortable VM environment
to Unix. The process continues today - there are thousands of hardware
and software developers using these core technologies:
I received a Distinguished Contribution Award for my efforts
on this project.
Postscript: I was told later that the results of our work
were well known to some of the venture capitalists in Silicon Valley. They're
amazed at the changes in the way Rochester behaves both in absolute and
in IBM relative terms. Some even claim that the change in Rochester's way
of doing technical work was the cause of the demise of DEC's VAXen. This
seems a bit strong, but it does provide validation that I was part of something
Distributed Debugger (1992-1993)
I was team leader for debugger which could debug code running on disparate
systems. This gave me the opportunity to do a lot of work in C++ and designing
objects. The work was tightly coupled to changes being made for new hardware
and software systems being done at the time. The environment was highly
volatile as both the requirements and hardware details changed on a weekly
In the early 1990s IBM had gone through several "downsizing" efforts
and I decided to take IBM up on it's generous offer. If they were going
to pay me money and provide other benefits, who was I to say "no".
So in July of 1993 I left IBM to see what the world had to offer.