Leaving IBM Mainframes Behind

Calma

 

As I mentioned above, when I decided to leave BofA I got a number of job offers. All of these jobs except Amdahl, entailed leaving behind working on IBM mainframes. (Technically Amdahl wasn’t IBM either, but it was IBM compatible.) All these companies used various kinds of mini-computers, an aspect of the computer business I had yet to be exposed to. I was nervous about leaving the realm of IBM, because I knew my IBM experience would rapidly become stale, and IBM was, after all, the dominant company in the computer business. (Many people in those days simply referred to all computers as “IBM machines”.)

 

I finally decided on Calma, a company that had started out making pen plotters. These were large devices which drew pictures by mechanically moving a pen over the surface of a sheet of paper. Since there weren’t any real application needs for such a device, Calma invented one: A Computer Aided Design (CAD) system to help designing integrated circuits. This product and its successor (GDS-II) became almost universal in the IC industry, with every major player, from Intel to NEC using Calma equipment. In fact while I worked for Calma, I spent two years in Tokyo, Japan, acting as a technical liaison to our Japanese partner, C. Itoh, Company. Our customers in Japan included all the major Japanese electronic companies: NEC, Hitachi, Matsushita, etc.

 

My first assignment at Calma was to learn the DEC VAX computer, and its operating system VMS. The VAX had an amazingly rich and complex machine language. I remember that you could compute a CRC with a single instruction. This was due to the fact that VAXen were microcode based machines. Not only did this allow various models of VAX to share a common machine language, as was true of the IBM System/360 and System/370, but it also allowed for arbitrarily complex machine instructions, by having underlying microcode routines of comparable complexity.

 

 

I loved the VMS operating system. It was well designed and intended to be interactive. Like CMS, there was a single integrated operator and user command language call DCL (Digital Command Language). It had an excellent online help system, which made it easy to use, and an extensive run-time library which made it easy to program.

 

I spent several years working for Calma. During that time I first worked for the CalmaNet group. At Calma I didn’t have a private office—I worked in a cubicle. I had my own DEC VT-100 terminal at my desk.

 

My second assignment was as an R&D liaison with a Japanese joint-venture company called GEIAL—the General Electric Industrial Automation Limited—in Tokyo. This assignment was to be for one year. I spent six months before I left getting ready, meeting people and familiarizing myself with other products in the company.

 

In Tokyo we ended up living in a luxurious apartment only one block from my office. When I first arrived at GEIAL I had no access to a computer. Eventually we bought several HP office computers—they ran MS-DOS but were not IBM compatible. I did no programming of these computers, although a Japanese woman that worked for me wrote a bug-tracking system in dBASE.

 

When I returned from Japan I went back to work in the CalmaNet group. This is where I first started working with the UNIX system. UNIX was the hot technology at the time and I was happy to work on it. UNIX had several unique characteristics at the time which have come to be more mainstream since then. It was written in the C language rather than assembly language. It was written to be portable, to run on more than one kind of hardware. And it had integrated networking allowing us to connect to another innovation of the time, the Internet.

 

Another trend was happening at the time, the advent of the personal computer. I’ll have more to say about the personal computer in the next chapter but I will have one thing to say about it here: its affect on Calma. Calma systems were based on mini-computers that cost in excess of $100,000. Once PCs became powerful enough—which they were doing by the late 1980’s—it became feasible to run CAD systems similar to Calma’s on a PC at a fraction of the hardware cost. The net effect was to make Calma systems obsolete. The company was too slow to recognize this technology change and didn’t survive.