Printer Friendly

Workstation wars.

Measured against the size of the mainstream personal computer market, high-end platforms--proprietary workstations, minicomputers, and file servers--are almost an afterthought: Sun, with a 33% share of the market, ships a mere 16,000 machines a month; Hewlett-Packard, with 21%, about 10,000. Nevertheless--as the new Apple-IBM alliance suggests--it's finally dawned on the major PC hardware vendors that the high-end market is more than a collection of exotic niches. There's a growing consensus that whoever controls workstation-class standards probably will end up dominating the next generation of personal computers as well. Clearly, that's the logic behind Apple and IBM's otherwise bizarre plan to spin off a new, jointly-owned software unit to develop an "open" object-oriented operating system. And the same jockeying for position underlies another odd partnership, the "Advanced Computing Environment" (ACE) consortium that Compaq, Digital, Microsoft and several dozen other spear-carriers announced two weeks earlier. The Patriot Partners venture (IBM likes to cover all its bets), Microsoft's promised 32-bit NT version of Windows, and lots of other recent non-aggression treaties all have a similar goal: To control the high ground for the next generation of high-performance hardware and system software. Alliances like ACE and the Apple-IBM deal almost invariably debut with a blare of headlines and macho talk about how much of the current market the participants already "control." The truth is, however, that both ACE and the Apple-IBM alliance are acts of desperation. Apple and IBM in particular have frittered away their best chances to become serious high-end players; now, we're asked to believe that two traditional rivals (the same folks who couldn't get OS/2 and System 7 out the door on schedule) can whip together a three-year campaign to produce the dominant operating system of the future--on three different processor architectures, no less. Sure thing, guys. Of course, it's interesting to note that most of the major hardware players in this race have come to the same conclusion: Software environments are going to define future platforms, not raw MIPS. In effect, hardware has now become a commodity even to hardware vendors; companies are beginning to compete on the basis of differences in system software. As this competition intensifies, a single generic operating system--like DOS or Windows--probably won't dominate the next-generation of personal computers. Instead, developers will almost certainly have to deal with a marketplace split among even more competing platform vendors than we now have. Thus, to the software community, the current workstation wars are more than just a spectator sport. The high end of the market is likely to undergo severe shakeouts before a few platforms establish themselves as dominant. At least one or two of the new platforms are going to open up important software opportunities; many are bound to be dead ends. But picking a likely winner at this point is still a tough call. The current front-runner is probably Sun, but Sun still has to prove that it can transform Unix into a state-of-the-art, object-oriented environment. The other major workstation and mini companies--Hewlett-Packard, Digital, Silicon Graphics, Intergraph--may be too closely identified with niche markets to transform themselves into industry leaders. Apple and IBM (not to mention little Next) at least understand that image is as critical as technology in fighting a standards war. How many workstation companies, after all, get coverage on national television just for announcing a joint development plan? In the long run, however, charisma and technology alone won't win the workstation wars. our guess is that there are at least three very tough issues that next-generation platform designers will have to face: * General-purpose or task-specific? Traditional workstations and

minis tend to be optimized for specific tasks--scientific and

technical applications, graphics, database management, document

creation, etc.--while personal computers have always been

general-purpose machines. This simple distinction has large

implications for software design, market dynamics, and computing

styles. (Users of general purpose machines tend to buy a variety

of stand-alone applications from multiple vendors; users of

specialized workstations tend to buy tightly integrated tools

from single vendors.)

The one significant exception is the NEXT machine, which

purports to be a true general-purpose workstation platform.

We're not persuaded that NEXT picked the right model; our guess

is that the really successful platforms--and system software--

will end up further along the task-specific spectrum. But can

committee efforts like the Apple-IBM or ACE alliances deliver a

well-focused design? The best prediction we've seen (courtesy of

New York Times columnist Peter Lewis) comes from a German-speaking

Hewlett-Packard executive, who predicts ACE will create

an "eier-legende Wollmilchsaull--an egg-laying woolen milk pig. * Backwards compatibility: Apple and IBM are already cranking out

press releases promising that "applications written for current

operating systems, including AIX, OS/2, and Macintosh, will run

in these new environments." That's probably a half-truth at

best, but let's ask the heretical question: Why bother? The fact

is, the best way to cripple the design of a new platform is to

lock in the limitations of prior systems. The Mac was a success

largely because it started with a clean slate; OS/2 and Windows

were crippled from the beginning by such archaic features as the

286 real mode and a file system inherited from CP/M. Of course,

brand-new platforms are born with not much software support,

which makes the marketing folks nervous. But users don't invest

in new platforms simply to run old applications. The real value

of a new platform is that it lets users do things that were

difficult--or even impossible--with earlier systems. * Delivery time: one reason most operating systems aren't finished

on time is that their designers get caught up in a kind of

technology megalomania. The new environment has to become an

engineering tour de force, embracing every sexy concept and

leaving rivals gasping in the dust. In the end, the job takes so

long that the system architecture becomes obsolete before the

first release is finished. The challenge for Apple, IBM, and

other would-be workstation competitors is to get a product into

the marketplace that delivers the right functionality on time.

In this market, an over-designed operating won't just be an

embarrassment--it may be fatal.
COPYRIGHT 1991 Soft-letter
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1991, Gale Group. All rights reserved. Gale Group is a Thomson Corporation Company.

Article Details
Printer friendly Cite/link Email Feedback
Date:Jul 7, 1991
Previous Article:Borland.
Next Article:The Gates memo.

Related Articles
Identity Crisis Jolts Workstation World.
Nonfiction Titles.
The Book of Ted: Epistles from an unrepentant redneck.
The Enterprise Team (Business Relationship Reengineering).
Weather forecasters train for perfect predictions.
Embodied love.

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters