In 1989 Houston was an oil town that had very little tech presence. When Compaq started (and quickly became huge), it was Houston’s darling. It had great benefits and everybody wanted to work there. They had a gorgeous campus built in the North West corner of Houston (not far from Tomball) that incorporated both office buildings and factories – all of which was connected by a second-floor enclosed walkway called “the spine.” You could walk between buildings without ever having to go outside.
They also owned a lake across the street that we could use for canoeing on weekends.
My job was in the “DOS development team” – DOS being the original Disk Operating System sold by Microsoft. Compaq products sold for a ridiculous premium, and part of the Compaq “value-add” was it’s own version of DOS that was optimized for Compaq products. Hence, we had licensing and source code agreements with Microsoft. PC clones were prevalent at this time, but they often had quirks, incompatibilities, or were hard to setup. Compaq’s business model was to remove this friction.
As I was an “application’s guy” rather than a “systems guy” (which composed most of the team), my first project was called the Owner Registration Diskette. It was essentially a marketing survey that buyers would run on a disk, their results were saved to disk, and then they’d mail the disk back where the results would be stored in a database (this is how things worked before the internet).
This is simple code to write, but the problem is that the marketing team wanted to update the survey very often, and nobody wanted to rely on engineers to make every update (especially the engineers). I was the only engineer on this project, so I first started by designing a front end that would allow the marketing team to make their own updates (which included very simple logic, like if they select choice C, select these other questions to ask). I decided on using a cross between a spreadsheet and a flowchart to represent this logic, and it worked very well (I even earned a patent from it). When they were done updating the logic in the front-end, they would press a button that generated a file. A second program would read the file, follow the logic, display the correct questions, and then record the answers on the disk.
It was kind of a fun first project and it gave me a lot of visibility (especially due to the patent). I was also the only engineer that interacted regularly with non-engineers (mostly marketing) and I was also one of the few engineers that got to travel (mostly to the company in Reston, VA that managed the database).
My second project was the Compaq EZ Help Diskette, which was a program that allowed you to browse through a library of product documentation. Today this would just be a simple website, and Microsoft Windows also had an electronic documentation feature that would do something similar (but not everybody used Windows at that time – DOS and OS/2 were still popular). After I finished the project, I did a deep dive on Microsoft’s documentation system and was surprised (and somewhat happy) to see that they were almost identical in most ways. Mine allowed them to write documents using Microsoft Word for Windows, and then I would convert that to a file that the program could read. I leveraged a bunch of Word features for creating tables of contents, hyperlinks, etc. and it turned out that Microsoft’s documentation system did the same things in almost the exact same way.
Compaq was a super fun place to work until 1991 when they fired Rod Canion (the beloved founder) and changed their business model (it was a necessary change, but it was very unexpected). Companies tired of paying Compaq prices when much cheaper PC clones were very prevalent. Instead of writing all of our own software in-house, we started bundling external products. Many people were laid off as a result, including members of my team. This made a ton of sense from a business perspective, but it was still a blow and morale dropped hard.
My last major project before leaving Compaq was to partner with Microsoft’s new Windows Sound System. Compaq helped create a new audio chip (which we embedded on certain Compaq motherboards, and the speakers were part of the keyboard), while Microsoft created an add-in sound board and some software for it (e.g. a sound mixer, sound recorder, plus some low-level audio drivers) which Microsoft sold as a retail product (it flopped due to the prevalence of Creative Lab’s SoundBlaster product).
Compaq called this feature “Compaq Business Audio,” and it also flopped (as most businesses people do not want their computers to talk out loud – Star Trek lied!) But the Compaq hardware had enough differences from the Microsoft hardware that all of Microsoft’s software had to be tweaked a little bit. This was my first real introduction to system program and to Windows programming.
By early 1993, Compaq was no longer a great place to work and everybody started looking for new jobs. My friend and co-worker Mike Yonker found a job with a small startup in Plano, TX named Pixel Semiconductor that made video chips. A few months later he asked if I’d be interested in joining him, so I had a few interviews, and eventually worked out an offer.
April 16th, 1993 was my last day at Compaq and it was hard. It was the only real job that I had known, the people were nice, and they had a huge, flashy facility. I was leaving to be the 26th employee at a tiny company in a crappy building in an ugly business park. I just hoped that everything would work out.
Overall, I worked for Compaq four years and three and a half months.
Here’s a good article on the history of Compaq: https://www.pcguide.com/news/what-happended-to-compaq/