I’ve been working in the computer industry since 1986 when I first worked at IBM in Boca Raton as co-operative education student. I was lucky enough to work for IBM in 1982 in Tucson, Arizona during high school senior year but working in Boca was a totally new experience. It was the first time completely on my own and I really enjoyed it. There was a huge degree of freedom being in Florida for the first time. After having lived in the desert since 1978, Florida was a water paradise. My assignment was for only six months and I made sure to return the following year for another six months.
Anyways, the point of the introduction is to give some context to what I’m about to say. Working at IBM Boca Raton (actually Delray Beach) meant working on things related to the PC and its software. By the time I had arrived, they had already released the AT and were working on the next generation. As a side note, the AT project team had a nickname of the “End of the World Gang”. I remember asking why it was called this and the answer I got was that they actually thought they were going to build a machine good enough for all time. Pretty cocky huh? There was a project after AT that didn’t make it out the door and it really was a sort of hybrid between the AT and what was to come next. I remember that a few people at IBM had them in their offices and found them to be quite useful since they had both the 3.5 and 5.25 floppies in them.
The next generation became what is known as the PS/2. I worked on writing test code for the ABIOS system. ABIOS was designed to remove the weaknesses of BIOS (hence being called Advanced BIOS) and to guarantee compatibility between different models of PS/2. In theory, this was a really good idea since it allowed for the same drivers to be used for all the different hardware platforms that supported ABIOS. IBM did indeed take advantage of this for OS/2. However, ABIOS was not as good as the replaceable driver model found in modern operating systems. ABIOS was not accepted by the industry even though the concept had some value. PS/2 really showed how much IBM wanted to re-gain monopoly rights on the PC family. IBM had invested millions in revamping and finding ways to lock out other vendors from producing the same thing without paying licensing fees. Microchannel and extensive patenting preventing other vendors from fully embracing the new platform. It didn’t help that Compaq had already come out with a 386 platform before IBM and that it supported the old adapter cards.
1987 was really the beginning of IBM’s doom with the PC industry. IBM boldly stuck with things that made it different without realizing that hardly anyone else was going to follow their lead. It would ultimately lead to IBM dropping all the things in the PC business that made it different.
It really wasn’t about the technology. It was about how IBM handled the “openness” of the new systems. Somehow they had forgotten that the openness of the original PC was what lead to its great success.
It wasn’t until PCI was invented years later that a decent replacement to Microchannel was created. Some of you might remember EISA. It was okay, but not great.
I’ll leave you with one last memory related to testing ABIOS. Another student had put together the framework and I was adding to it the ability to test the individual ABIOS components. After a number of attempts on individual components, it became possible to test multiple components at the same time (disk, diskette, mouse, keyboard, screen, etc..). The first time I ran this test was quite exciting. Not because it was running but rather because it really was doing all these different actions at the same time. After experiencing DOS, this was my first view of a real multitasking PC system. Even more amazing, it worked pretty well.
After I returned from this first session at IBM, I decided to focus on operating systems at school. It was probably my favorite class in computer science at the University of Arizona.
In 1989, I went to work at IBM Boca for OS/2. This lead to Citrix in 1993.