Opinion: Greg Scott Mulls Alpha, HP and Antitrust Appeals

ENT columnist and DEC veteran Greg Scott considers the momentous summer that saw Compaq dump Alpha, HP offer to buy Compaq, and Microsoft's apparent dodge of a break-up.

By Greg Scott

Whew -– what a summer! First Compaq announced they are dumping Alpha, Microsoft won a partial victory in the appeals court, and now HP wants to buy Compaq.

By now, the Alpha announcement is well-known and almost ancient history. Compaq announced it would produce one more version of the Alpha chip, dubbed EV7, and then no more. Instead, Compaq says it will port its Tandem, OpenVMS, and Tru64 Unix operating systems to IA-64, Intel’s 64 bit architecture. According to the announcements, Compaq plans to get all this done in the next three years.

I can’t think of many tougher engineering challenges, especially since I understand that Intel owns the compilers this time, and I would not want to think about the marketing challenge Compaq will face.

I still remember the meetings I attended way back in 1990, where the leaders at Digital Equipment Corporation said that Alpha would set the standard for 64-bit computing. I believed all of it back then, and I still wear my black Alpha cap every once in a while. Perhaps it will become a collector’s item someday.

I also remember 1998, just before the $9 billion Compaq buyout, when Digital settled the big lawsuit with Intel. Intel paid Digital $750 million for the semiconductor plant in Hudson, Ma., and took over all chip manufacturing. I made some people mad when I said that Digital had just dumped Alpha. I don’t often get an opportunity to say, “I told you so,” but I can’t help myself this time.

I’m not happy about the decision to drop Alpha. In fact, I’m downright mad because Bob Palmer and a bunch of Digital vice presidents squandered an incredible technology advantage throughout the 1990s. But given the sad state in which these clowns left the company, I must grudgingly admit, the choice makes sense today.

So let us lift our glasses one more time and toast what could have been, what should have been, but what wasn’t.

Meanwhile, life goes on. By now, the whole world knows Microsoft dodged another bullet and will not be broken up. I didn’t do so well on my predictions for this one. I thought the legal system would barbecue Microsoft and I said nothing good would come from Microsoft’s tactics at the trial.

I was wrong. It seems that if defense witnesses contradict themselves, get caught in outright lies, and make the trial judge so mad that he blows his objectivity, they can still win because they provoked the trial judge into a bunch of dumb moves. I’ll bet the law schools don’t teach that strategy!

And to cap off a long, hot summer, HP now wants to buy Compaq. This blows my mind. Imagine the combined company with competing Intel server and PC lines, competing RISC server and workstation lines, and enough incompatible operating systems and variations of UNIX to fill an acronym dictionary. Imagine the politics trying to sort out that mess! So what does all this mean?

First, if I were in charge of Compaq OpenVMS or Tru64, I would look for another job. Even if the buyout doesn’t come to fruition, OpenVMS and Tru64 are in big trouble. Tandem has the reputation and market share in non-stop computing, Intel servers own the low end, and UNIX owns the midrange. Alpha is history and HP bet on IA-64 years ago when it partnered with Intel. Compaq Tru64 never achieved any significant market share, OpenVMS is slowly dying, and I have a hard time believing any serious IA-64 porting work is going on. That means the combined company will likely offer Windows 2000/XP, Linux, and maybe Netware on the low end, some variation of HP-UX with a few Tru64 porting tools in the middle, and Tandem at the high end.

If I were an OpenVMS or Tru64 customer, I would stay up at night worrying about the future of my applications and what platform to use. The world will not sit still and wait for HP/Compaq to sort itself out, and any customer who still trusts what’s left of the former DEC brain trust is nuts. Frankly, I would not believe any nondisclosure presentations from Compaq reps or their resellers unless they can show tangible evidence of real, ongoing engineering work and somehow demonstrate a commitment to continued engineering even after a buyout. And Compaq can’t demonstrate such a commitment because its leaders want to sell out.

OpenVMS and/or Tru64 could spin off into its own company. Here is a marketing niche an OpenVMS company might exploit: Just recently, I had a heart-to-heart conversation with a long-time customer complaining about forced upgrades with Microsoft products. She pointed to an OpenVMS Alphaserver in the data center and said it just works, year after year. (I should know; I set it up.) Then she pointed to yet another Intel server we were installing and complained that her operation now depends on a complex web of these servers, and if any one of them fails, her data center has big problems.

She’s right. Given recent history, I wonder if a mass market now exists that cares about reliability?

Given enough financial backing and a leadership team with enough guts, an OpenVMS company might work. But the problem is, who would supply the money and the guts?

Meanwhile, Microsoft rolls on. I have criticized Microsoft’s business practices and blatant Windows bugs in this column, but I will pay Microsoft this compliment: Microsoft’s leaders appear to be absolutely, rock-solid, committed to Windows for the long-term future and Microsoft, unlike its competitors and hardware partners, has so far resisted the temptation to constantly panic and grab on to an ever-changing strategy du-jour.

How will it all turn out? Who knows? All I can say is, it won’t be boring to watch.

About the Author

Greg Scott, Microsoft Certified Systems Engineer (MCSE), is founder and Chief Technology Officer of InfraSupport Etc. Corp. (Eagan, Minn.).