THE WARPED PERSPECTIVE
May 1999

Does anybody really need a new OS/2 Warp client version?

Does anybody really *want* such a thing?

Is anybody willing to PAY for it?

Well, sure, I'd buy a copy. I'd buy one for my parents' computer, too. Another one for my machine at work. I have two consultants panting to get their hands on a copy at work, as well as two or three other customers. Of course, then there's the Boss's machine, running Windows 3.1.

Dollars to doughnuts, you are in that same position, right?

But even if everybody who reads this article buys ten copies, that would only amount to around 50,000 copies or so. And if each of us represents nine other OS/2 users who don't read *THE WARPED PERSPECTIVE* each month (yet!), that would still only total around 500,000 copies of an OS/2 Warp 5.0 client. If we add in IBM's business clients who want it, that total can go up by a factor of ten. But that's still only five million copies of an OS/2 Warp 5.0 client.

Now you and I, we're no fools. You and I could sit down with a scratch pad and figure on a tidy little profit from selling five million of *anything* -- even toothpicks! Even IBM could make an immediate, short-term profit on the sale of five million units of product, despite their overhead. The problem -- the *real*, honest-to-goodness, down-to-earth, IBM-class problem -- is that this event would only be the very first step in a long list of events that would then transpire. Let's list a few of those events.

First, a product announcement would have to take place. IBM could obviously hide that underneath a Linux or NT product announcement, but they still would have to deal with assigning SKU numbers, release dates, shipping schedules, etc. IBM would have to spend money on package design, production costs, storage costs, etc. For you and me these costs would be trivial, but not for IBM.

Next, a support mechanism would have to be assembled. IBM must provide support for its products, because IBM customers demand and expect it. IBM cannot simply cut off support for an older version (the way Microsoft likes to do), because IBM is making revenue on these customers who still use the older version. These customers, particularly large banking, insurance, and manufacturing corporations, are even more slow-moving than IBM itself. They cannot and will not dump a good, working product like OS/2 Warp 3.0, particularly with FP40 including all necessary Y2K fixes. So IBM's support overhead for OS/2 would rise signficantly if they added a third version to the mix.

Finally, the inevitable product compatibility issues would arise, and associated fixpacks would have to be produced. IBM's slow-moving customers expect to see continued OS/2 3.0 fixpacks for the next few years, at least, and Warp 4 fixpacks for ten or twenty more years. A new set of fixpacks for a Warp 5 client would add a signficant burden to IBM, which can't simply abandon the Warp 3 customers the way Microsoft has abandoned its Windows 3.X customers of the same era. This third OS/2 version under support would also add to the support costs of IBM customers who would have to then deal with three client versions instead of just two.

You see, software is very unlike toothpicks, mouthwash, or other physical products. These products are either consumed or wear out. They don't develop a huge infrastructure of other devices that rely on them, as software does. Toothpicks don't have support costs and fixpacks and compatibility issues and device drivers. Throwing away a toothpick, or getting rid of an old jalopy, or even moving to a new house -- these events don't have ripple effects on what kind of food you can eat at your next meal, or what kind of car stereo you can buy, or what kind of clothing you can wear. But information systems are too tightly linked to allow for casually tossing aside one product and plugging in a different one.

Certainly a great deal of this anguish is caused by Microsoft's preload monopoly. Read the license agreement for Windows98 sometime. Did you know that support costs must be borne by the *PC manufacturer* instead of Microsoft?? It's in the contract. Do you think that having a 20% preload market share would lighten IBM's support load by encouraging the growth of third-party support companies? You bet your sweet rhubarb. The problem is that IBM's fixed cost per client version is not going to be counterbalanced by a large influx of new OS/2 customers. The variable (sales) part of the equation does not exceed the fixed (development & support) part. It *does* do that if you're a monopolist. Call it Softonomics 101.

We are fortunate indeed that IBM is big enough to keep supporting two distinct client versions of OS/2 already -- version 3 and version 4. We will likely have to wait until version 3 rides off into the sunset to see a version 5 that is a signficant step-up in terms of power, features, and enhancements. Alternatively, perhaps a dedicated team of OS/2 supporters could assemble a long-term support infrastructure that would ease IBM's overhead burden. But who has the kind of capital necessary to develop such an arrangement? Not the you and I who just want to buy ten copies of OS/2 Warp 5.0 apiece.

Is there anything we OS/2 users can do to help IBM get their act together? The key will be to develop OS/2-based business solutions that generate signficant capital that can be used as leverage to get what we want from IBM. (Check out http://www.serenity-systems.com/, for example). The only other possible solution is to write some kind of "killer app" that will entice millions of people to clamor for OS/2 on their PCs. Pick your poison.


Most recent revision: April 28, 1999
Copyright © 1999, Tom Nadeau
All Rights Reserved.

E-MAIL: os2headquarters@mindspring.com