PCjs Machines

PCjs Blog

A COMPAQ DeskPro 386 Mystery Solved

On September 9, 1986, at 11am at the Palladium nightclub in New York, COMPAQ unveiled the DeskPro 386, described by COMPAQ President Rod Canion as “the most advanced high-performance personal computer in the world.” With this product, COMPAQ effectively declared that they – along with Intel and Microsoft – were now “setting the pace for the rest of the industry”, by creating “the first of a new generation of industry-standard desktop workstations.” *

It was a major announcement. Speakers included Ben Rosen (COMPAQ Chairman), Rod Canion (COMPAQ President), Bill Gates (Microsoft Chairman and CEO), and Gordon Moore (Intel Chairman), followed by software industry leaders from Ashton-Tate and Lotus.

However, missing from the media spotlight were any of the engineers who worked behind the scenes to make the DeskPro 386 a reality. Even Rod Canion’s book “Open: How Compaq Ended IBM’s PC Domination and Helped Invent Modern Computing” doesn’t mention many engineers. He does say that “Every part of the company had contributed to [the DeskPro 386’s] success, and many dedicated people had worked long hours in the summer months leading up to it”, but his focus was almost exclusively on upper management.

Presumably many of those DeskPro 386 engineers also appear in this company photo, which is undated, but assuming the machines pictured are COMPAQ SystemPros, it’s probably from 1989 – roughly 3 years after the DeskPro 386 was introduced.

As for the DeskPro 386, it appeared that some of the engineers had their initials “inscribed” inside the machine – specifically, inside the DeskPro 386 ROM – but it was unclear who CAB, GLB, RWS, DJC, and NPB were.

A COMPAQ Mini-Reunion

Last year, HPE firmware engineer Thomas Palmer (who still works at the location of COMPAQ’s old headquarters in Houston, Texas) took an interest in this part of COMPAQ’s history, tracked down as many former COMPAQ engineers still in the area as he could find, and made arrangements for a small informal reunion.

The reunion happened one year ago, on June 19, 2018. Thanks to Thomas, I was able to attend the reunion, and the engineers who worked on the DeskPro 386 were gracious enough to sign my copy of the COMPAQ DeskPro 386 Technical Reference Guide – and in the process, solved the mystery of the DeskPro 386 ROM’s principal authors:

• Monty McGraw
• Mike Perez
• Carrie Boone (CAB)
• Norm Brown (NPB)
• Randy Spurlock (RWS)
• Darren Cepulis (DJC)

Of the 5 people whose initials are embedded in the original ROM, only GLB was unable to attend.

When Mike Perez noticed my copy of the TRG, he told me a bit more about it:

MIKE: What is that?

JEFF: This is the Technical Reference Guide for the DeskPro 386.

MIKE: You know who wrote that?

JEFF: No.

MIKE: Me.

JEFF: Really? I’m hoping to get it autographed actually.

MIKE: Because we had a terrible writer that was taking way too long…. So Kevin Elton said, no, we got to have this thing done by announcement.

So I would write one chapter one night, review that chapter the next day, while I wrote the next one. And then it was all done and they forgot to order the cover–the binder. So the binder was the last item that came in. And Kent Price–rest in peace, he just passed away about maybe a month ago–he was in charge of the documentation and we would go there and rake him through the coals.

The most controversial part of that book was the errata, because we had to work with Intel. There was some errata that we fixed in code or in hardware… so I put it in there that it was fixed.

JEFF: Usually Intel was very tight-lipped about the errata, so I’m surprised they let you publish it.

MIKE: That’s what was controversial about it. So the only errata we put in there was the couple of errata that were not fixed. There was a floating point issue, if I remember correctly…. Backtracking a little bit: we had bought, we had already bought the first lot of 386s coming from Intel. We bought the whole thing, sight unseen. And so our job was to make it work, despite all the errata.

And so we we did a bunch of fixes, some in hardware, some in the ROM, to get around the errata. So the only errata that were left was like the floating point one…. And so, there was a bunch of errata that we didn’t publish… good luck to the people who were trying to do their own. You gotta find them on your own and fix them on your own. The other little tricky thing is we asked the hardware… the key hardware designer to leave a couple parts out of the schematic just in case somebody took the schematic and copied it.

CARRIE: COMPAQ was so associated with the 386 that that’s why Intel started their “Intel Inside” campaign, because people thought that COMPAQ had created the microprocessor.

Some of the engineers are now retired, like Randy (who also worked at NVIDIA), while others are still working in the area. Monty still works at HPE and Norm still works at HP Inc. And others have moved on to other careers, like Carrie, who is now a patent lawyer at KDB.

Memorabilia

Several of them also brought early COMPAQ memorabilia to share, like this “Magnum” t-shirt. I learned that Magnum was the code name for the DeskPro 386 (one of the little details Canion neglected to mention in his book). No one would probably put imagery like that on a company t-shirt today, but those were simpler times.

One of them donated this jacket to me, and the “COMPAQ/Microsoft” baseball cap – a gift from Thomas Palmer – completed the fashionable ensemble.

Another engineer brought a collection of COMPAQ newsletters, which I hope get scanned and preserved online someday.

Thanks for the Memories

I got the sense that everyone really enjoyed the opportunity to reconnect and reminisce. Thanks again to Thomas for taking the initiative to organize it, and for giving me the chance to participate and show appreciation for all their hard work.

References

“Open: How Compaq Ended IBM’s PC Domination and Helped Invent Modern Computing”, by Rod Canion (p. 98). BenBella Books, Inc.. Kindle Edition.

[GitHub Source]

Jeff Parsons
Jun 18, 2019