An American Programmer
at a French Computer Maker - 1965-1966

Computer History Vignettes

By Bob Bemer

Following my stint at Univac, I followed my mentor Dr. Louis Rader to Bull General Electric, as he had need of loyalty and technical competence when GE purchased the components of both Compagnie Bull (Paris) and Olivetti (Milan and surrounding). Italy was easier -- they had no Charles de Gaulle to stir up dislike of Americans. So I was assigned to Paris, under Brainerd Fancher, a GE manager's manager with scant knowledge of the computer business, likeable though he was.

It proved to be a remarkable adventure and thorough re-education (not only in the language spoken). These are some of the happenings.

Headquarters

Bull was then at 94 Avenue Gambetta, in the 19th Arrondissement. A sprawling building for which they had no public map. A strange thing to an American was that they depended mostly upon daylight for lighting. I got used to working that way, and still do when possible.

Much caste-consciousness prevailed, depending largely upon the school one had attended. If the Ecole Polytéchnique, then you certainly ate lunch in a room reserved for "cadres". If not, at a much larger room for non-cadres. But both served wine.

I can't say that management was modern. When I first arrived, there were no dispensing machines whatever in the building. Then an old personnel type tried to fire one of the programmers that was seen having coffee at 10 in the morning at a nearby cafe. I charged into the man's office and irately told him that the reason was that he had been up all night running a problem on a computer, and I damn well wanted to see coffee machines soon, together with his reinstatement. I knew the man personally and he was too good to lose.

The machines were ordered, but the man was still irate and wanted to quit anyway. So I asked where the non-cadre lunchroom was, went there, and had lunch with him. Probably a terrible breach of protocol, and it certainly had tongues wagging with this precedent, but it worked. I persuaded him to stay.

Management

At first I was given a sort of roving authority, reporting to all of Pierre Davous (Engineering, and a wonderful man), Maurice Teper, and Andre Chargeraux (Marketing). Later I was assigned directly to management of software, where the rising star was Francois Sallé.

Software Production

I had come from UNIVAC, where I installed a labor distribution system -- the first for software, as far as I know. Managers were charged to estimate the size and content of software components, begin and end dates for construction, etc. On punch cards, signed. When changes became inevitable, for whatever reason, they were required to submit modifications, and sign again. If they failed, and caused their management to be deluded, they were going back to being plain programmers again.

But at Bull they tried to confuse me about completion dates. Nowhere was there a description of what the software was supposed to do, how it would do that, and by when. There was a lot of foot-dragging, and the local management could not see what I was trying to do. In desperation one day, I went to central Paris and purchased (with my own money) a huge schedule board, and strapped it to the back rack of my little Sunbeam Alpine. Later I put the cost on my expense account, which confused the typical French accountants.

DOS

In addition to new hardware and software development, Bull GE also sold the products fron GE in the United States, including the 400 and 600 lines. But the problem with the 400 was that the operating system was based on magnetic tape, seriously depleting its possible performance. Charlie Lecht and Ralph Stout of ACT (Advanced Computer Techniques, in New York, to whom I had given their first contract, while at UNIVAC), were touting a disk operating system for the 400, but could not get anywhere with the folks at Phoenix.

I figured I had an independent responsibility to Bull GE sales, quite apart from how Phoenix people felt, so I met sort of surreptitiously with Lecht and Stout in London in early January of 1966. I knew I was doing the right thing when I found a cartoon strip about future space activities in an English paper on the plane, where the hero said "Come. I will take you to the disc". I took it to our meeting, where it was adjudged a superb omen. I told them to start development.

On Feb 20 of 1966, I brought the Phoenix product planner, Dave Booth, to another London meeting with Lecht and Stout, who helped persuade him to not object to Bull GE funding it. Agreed, it made me unpopular with the Phoenix people I would work with in ensuing years, but in the end they had to agree that my action saved the life of the 400.

Meetings

A meeting I remember well was when I went to the Olivetti component in Milan. Following on IBM's 360 precedent, BGE was planning a family of computers. The French were to build the Gamma 140, while the Italians were to make the smaller Gamma 115. Of course they were to be compatible in every way. But I had called a technical meeting just to make sure. Before the meeting I was taken to a sumptuous lunch by Lionello Cantoni, who headed their software efforts. The aim was to soften me by food and wine so I would not be any trouble to their plans. There I first met Tony Pizzarello, with whom I would later be a good friend in Phoenix.

At the meeting I was introduced to Marisa Bellisario, head of product planning. I thought her a single lady, but learned later that Sr. Cantoni was her husband. The crux came when I asked for a comparison of the two instruction sets and character repertoires. Why did I want that, la Bellisario asked? To prove that the two machines would be compatible and run the same programs with the same answers, I replied.

"But we agreed last May that they would be compatible", she said. To which I responded "Marisa, that was probably the same time in May that my wife and I agreed we would be millionaires by today. But we aren't -- it takes more than agreement!"

So to force the issue (for which the Italians complained loudly to Dr. Rader) I set up the programming and usage manuals to be common to both machines. But when I looked at what each had so far I was so discouraged that I hired Miss Dorothy Walsh (Computer Usage Company in New York), who had worked on my PRINT I manual at IBM, bringing her to Europe for whatever time it took.

I supplied her with a Bull programmer, whom I sent to Milan to validate conditions on machines actually running. After a while he asked me which serial numbered machine he should provide data for. "They're all different!" Suspicions confirmed, and Bob Curry (advisor to Dr. Rader, whom I knew from COBOL days when he was with Southern Railway) wrote him a nice letter exonerating me.

It turned out well for Miss Walsh, too. I met her again while speaking at the University of Pavia (Italy), and she was happily married to an Italian. And Marisa went on to be President of ItalTel, the Italian telephone company. Upon her death they named a prize, for women of achievement, after her.

Another well-remembered meeting was the one on 1965 November, held by John Diebold Associates in Bad Godesburg (near Bonn). He had taken over an entire castle, and coopted the local marching band to greet the delegates. Everything was of the fanciest, and at the formal dinner the speaker was the United States ambassador to West (at that time) Germany. At the table where I sat at the head there was a man taking notes throughout. After the ambassador finished, a woman stood up and repeated the entire speech in German. Then our man stood, and did the same in French. Except that he interpolated and modified it considerably, which had everyone that knew French holding their sides to contain the laughter.

One permanent memory is of a Diebold panel where a U.S. person suggested that the verbs of COBOL should be translated to other languages, and the compilers be modified to accept them. To which a German objected quite strenuously. He noted that the different tongues were like italics or bold to set them off and make statements more readable. Note that this was prior to HTML.

Back to History Index          Back to Home Page