1983: Oh No! They’re Developing Their Own Compiler

Late summer, 1983. I asked Philippe how his software project was going.

“They are writing a new compiler,” he answered. “They’ve decided there is no good PASCAL compiler for MS-DOS. So they are writing a new compiler first.” The “they” was a team of Danish programmers.

Oh no! Seemed like bad news

In that instant, I imagined myself as waving goodbye to money that had sprouted wings and was flying away. So you understand, the idea of “they are writing a compiler first” sounded something like having to climb Mount Everest first. They were supposed to take some software that ran in one operating system and convert it to a second operating system. Instead, they were recreating an entire programming language.

I was sure this was disaster. I had skin in the game. I’d foregone my consulting fees with Philippe and had taken one percent ownership in the company — Borland International — instead. And yet, in fact, it was the key to success. It turned out to be a spectacularly good move. The compiler they wrote became Turbo Pascal. Look it up. In fact, three links for you: 1.) https://en.wikipedia.org/wiki/Philippe_Kahn … 2.) https://en.wikipedia.org/wiki/Borland … 3.) https://en.wikipedia.org/wiki/Turbo_Pascal

The project in question was converting some menu interface software that ran on one existing personal computer operating system (CP/M if you’re interested in history) to a newer more marketable operating system (MS-DOS). You have source code written in PASCAL, a programming language. The PASCAL compiler reads the code and converts it to machine language, so that any computer using that operating system can run it as an application, a finished software product. To move that software application to another operating system, you start with the source code as a text document, load it into a PASCAL compiler built for the other operating system, and run it there (compile it) to create machine code that works in the new operating system. It’s called “porting” software.

In 1983, porting software was all the rage.

Personal computing was still in its infancy. There were a million or so early Apple I and Apple II computers around, and the Apple Lisa (look it up) just started. There were a single- million personal computers in businesses. One source says there were fewer than a million home computers, another says single-digit million microcomputers being used in business. Business computing on the new microcomputers (aka personal computers) was a huge growth market.

Philippe was Philippe Kahn, a software pioneer, a tall, portly Frenchman in his with an overabundance of gestures, enthusiasm, optimism, and drive. He had founded Borland the previous May. You can look him up too. He became a celebrity entrepreneur and was later credited with inventing the camera phone. He was about my age then, early thirties. I had done his business plan, and redone it, and I’d kept on revising almost weekly. He was not the easiest person to deal with, a bit more certain than I was used to, but I liked him and liked working with him. He was originally supposed to pay me a couple of thousand dollars for my consulting, but as time went on and my fees accumulated, he asked me to take equity and a seat on the board of directors instead of fees. I’m not sure I really had a choice, but I did take him up on the offer, so I became what we called then “a co-founder” of Borland International.

That moment with Philippe became on of those key moments, engraved in memory, for which you even remember where you were at the time. We were in the parking lot of the office building at Keily Blvd and Saratoga Ave. in San Jose, where I had my offices with Creative Strategies International, and Philippe had a rented single desk in some other software business. We had met for lunch and we were walking and talking on our way.

I’d fallen in love with personal computer software

As this happened, I’d fallen in love with personal computer software and programming, and made a career of it. I discovered programming with the BASIC programming language on the Tops 20 minicomputer in the basement of the Stanford business school. The admittance packet included a book titled “Teach Yourself BASIC” and I did. I became a paid monitor in the computer facility, helping other students with programming. I got my part-time employer, Creative Strategies, to let me program a forecasting system on their computer as a summer job. Later, while I went full time with Creative Strategies after the MBA, I wrote an entire bookkeeping system in CBASIC, for a consulting client. Then I started a group within Creative Strategies to do market analysis and market forecasts of that business. I became the VP in charge of micro (personal) computer software market analysis. Which is where Philippe found me, with the recommendation of my older brother, who was his lawyer.

The amazing board meeting

Two months later, early October of 1983, Philippe called a board meeting on a Saturday morning in his new offices in Scott’s Valley. Turbo Pascal was ready to go. It was sensational software, way better than the $440 Microsoft Pascal then available for MS-DOS. To give you the idea, if you were working with the existing PASCAL language, when you had a program ready to go you would tell it to compile and wait 20, 30, or more minutes. Programmers would set the compile and go to lunch. And, most of the time, the compile command would just stop with an error message, but not tell the programmers what error and in what line of code.

What we saw that October morning in 1983 was sensational. You’d manage the lines of code like normal. But when you told it to compile, it would start going and almost immediately, seemingly instantly, kick out an error message with the first error encounter, plus the line of code it was in, and what was the error message. It made programming in PASCAL 100 times easier.

Philippe’s business genius

Philippe seized the moment. I said he should price Turbo Pascal at $500 because it was way better than Microsoft’s $440 PASCAL; and price is the best marketing message. Philippe said no, we would price it at $49.95 and blow up the market. They would cost us only $5 to build (there was no Internet and no downloading so selling software was selling a disc and a manual. Philippe was right. We created a sensation. Philippe bet the business on a few full-page ads in the main personal computer magazines of the day. That worked. Turbo Pascal took off. Suddenly, almost overnight, Borland was a big deal. Everybody wanted Turbo Pascal. Sales poured in. Philippe hired people and Borland was up and running. And I am proud I was a part of it, although I don’t claim much credit. Philippe made the call against my advice. At least I had the sense to recognize sensational software when I saw it.