UNIVAC - and Computer Sciences Corporation

Computer History Vignettes

By Bob Bemer

When I arrived at UNIVAC to take charge of their software operations, I inherited one C. L. McCarty, who had rather wanted the job himself. He had made some preliminary negotiations with CSC to do the software suites for both the 1107 (Minneapolis-St. Paul) and UNIVAC III (Blue Bell). Approving or disapproving such contracts was an early and priority task. Before I could get out to Los Angeles, however, I fell out of a tree at my Weston home while trying to saw off limbs with tent caterpillar infestation and broke several lumbar processes.

But the decision could not wait, and as soon as I could get discharged I was on a plane to Los Angeles with many pain pills. That was on May 27 of 1962. I went to the CSC offices, which were then in a small building on the Palos Verdes peninsula. Presentations were made by Joel Erdwinn re operating systems (EXEC II), Lou Gatt (formerly Los Alamos) for FORTRAN, and Dave Ferguson re assemblers. I was much taken by the PROC capability that Ferguson presented, and it was probably in view of that (plus my knowledge of insularity in both UNIVAC software locations, and prior acquaintance with most CSC personnel) that I approved a contract. This despite the fact that I had once fired Fletcher Jones, the millionaire chairman of CSC.

I caught much flak from Univac Marketing people at first, at least until the software was operable. But it had a slow start. I kept a suitcase with a set of clothes in my New York office (Sperry Rand Building on the Avenue of the Americas) for fast emergency trips to Los Angeles.

Our major problem was that CSC management thought they had UNIVAC in their pocket, and could take their time making the software. Maybe they assumed it because I knew all three of the principals -- Fletcher Jones, Roy Nutt, (and Bob Patrick at the start). And they were also doing the software for Philco computers, whose local representative, Caral Samson (now Giammo), we called "Miss Philco".

It finally got so desperate that UNIVAC ordered me to take up residence at CSC to control the work. At first I went for a single month (from 1963 Mar 08 to Apr 02), and then for a 3-month period (from 1963 Jul 02 to Oct 08).

It was in the second sojourn that I found my leverage. Our Joel Boardman, reporting to C. L. McCarty, was in charge of our local liaison operation at CSC. Boardman kept complaining about his salary, as though it was more expensive to live in LA than New York. He got me seriously concerned about this, so I went to the real estate office that had rented him a house in the Palos Verdes area. Was he behind in his rent? The agent said "Oh no, the rent has been paid a year in advance!" "How could that be?", I asked. "Oh, it was paid by a check from Computer Sciences Corporation!"

Jackpot! That could cause a lot of slowness. Two things resulted. First, I fired Boardman on the spot, He complained to the California Commission concerned with this, which queried me. I told them that he had behaved improperly as a manager; this was a good enough reason for them. The second was that UNIVAC now had a large lever to pry work from CSC. The key was that they were about to go public with a stock offering! They couldn't afford to have a scandal known.

We confronted Fletcher Jones. UNIVAC VP Lee Johnson flew out, as did Milt Bryce of my operation, to help audit. UNIVAC demanded, and got, access to their books, and a written commitment of certain percentages of each of the key programmers' time to be applied to UNIVAC software fabrication. We signed up x% of the time of Erdwinn, Gatt, Ferguson, Owen Mock, and others.

Even then it wasn't easy going. We were still behind. John Butler, caring for our interests at U. S. Steel, had been pestering me for completion of Gatt's 1107 FORTRAN processor. When it was finally doing well, I called Butler very late one night from the machine room, saying "Listen to FORTRAN run", and put the telephone to the printer recording the compilation. His wrath subsided.

That FORTRAN processor was so good that IBM later let a contract to Ascher Opler of CUC (Computer Usage Corp.) to see why Gatt had succeeded so well. In a talk to FORTRAN Pioneers, Fran Allen of IBM, who later made a scholarly study of software processors (under John Cocke, who invented RISC processors) said it was due to bidirectional tape read. Eric Clamons and I thought it was clever use of the drum, which showed up at the right place just as needed. I believed that so strongly that I inspected the makeup of the first seven 1107 systems that had gone out. NONE had a drum! So I made an edict that ALL 1107 systems were to have a FASTRAN drum to make a good match with the software. The salespeople grumbled mightily until they saw how much the performance improved.

When the FORTRAN processor got to the field, a FORTRAN session at a UNIVAC Scientific Exchange (USE) meeting heard some complaints that could have led to a sales flap, so I offered a free diagnostic service for user programs evincing apparent bugs. One(!) was sent in; our reply showed their misuse of the language. I was always proud of this device to deflect customer complaints.

The CSC experience held many good moments. Wanting a Model 33 Teletype to communicate with my New York office, I called Pacific Telephone and Telegraph one morning to order one. They said I could not get one under any circumstances in less than three months. Wanting it now, I called John Auwaerter, V. P. Engineering for Teletype Corporation, creator of that series of equipment, who had led the X3 Committee in making ASCII a standard. Hearing my story, he said that he had just had lunch with AT&T's Vice President of Marketing, and would see what he could do. About two hours later I was besieged by calls from PT&T. My Model 33 was delivered the next morning!

Fletcher Jones was also interesting. He was a high flier in more than one way. And he had a thing for airlines stewardesses, especially those of TWA. While we were still in the smoke and mirrors period he invited my wife and me to his (rented) estate on the Palo Verde peninsula. A department store magnate and his wife were there, and Fletcher had the usual stewardess, who doubled as kitchen help. I said later that I was the only non-millionaire there.

About the shenanigans, Roy Nutt or Bob Patrick would never have been involved. Very straight shooters. Patrick had probably left by that time, anyway.

Back to History Index            Back to Home Page