February 2000

Raise your hands if you know the answer: What is the purpose of "Revision Numbers" or "Version Numbers" of software applications?

The short answer is: Marketing. That's it, plain and simple. The purpose of changing a version number is to imply to potential users of the software product that there has been some kind of significant improvement in the software. Note that the versioning does not "legally" make this claim; nor does such a version number change "prove" that anything inside is different. But version numbers for software serve the same purpose as Model Year numbers on cars: they imply newness without necessarily delivering it.

Probably the most obvious example of this kind of chicanery occurred in the early 1990s when Microsoft decided that it could not catch up with WordPerfect technologically, so they decided to catch up numerically. Microsoft named the version that followed MS Word 2.0 for Windows as -- MS Word 6.0 for Windows!!! That's correct, you will never find a copy of Word 3, Word 4, or Word 5 for Windows. By skipping three version numbers, Microsoft attempted to imply that MS Word 6.0 for Windows was an equal or perhaps superior competitor to WordPerfect 6.0 for Windows. (Considering that WordPerfect had jumped straight from zero to 5.1 in its first release of a Windows version, perhaps Microsoft's marketing response is not surprising.) Whatever you may think about the morality of such product labeling, this example shows clearly that versioning is primarily a selling tool, not a technological explanation of things.

Since we therefore cannot rely on versioning to guide us as to the relative technical merits of various software products, what about the Model Year numbering system adopted by Microsoft and others starting in the mid-1990s? Is it possible to conclude sight-unseen that Lotus Smartsuite97 is less powerful than, say, MS Office 2000? No, it is not possible to categorically make that assumption with any accuracy or truthfulness. However, most members of the buying public will make that assumption -- and Microsoft incurs no legal liability for this situation, because they did not have to make any bold claims about product superiority. They just let the numbers make their subtle, hidden suggestions. Isn't marketing wonderful? <gag!>

With OS/2 Warp we are in a similar situation. Until this month, Microsoft has released only one "new" version of an operating system (the "newer" product known as Windows98) since the 1996 release of OS/2 Warp 4.0. Of course, we know very well that Windows98 is a DOS system with a Windows shell, and is therefore an older and less powerful technology than OS/2. So it should not seem that OS/2 is "behind" or "old" compared to Microsoft products. OS/2 v4 is two months newer than WindowsNT v4, for example. Yet Microsoft makes a big deal about slight adjustments to its operating systems, using "OSR2" (OEM Service Release 2) of Windows95 as if it was a new and different product designation instead of a minor patch. (Notice the clever confusion that Microsoft caused by using the OSR2 name, in order to hide the name OS/2 by misassociation!)

This month, another such marketing trick is being played on the buying public. By renaming the update to WindowsNT 4.0 as "Windows2000", Microsoft hopes to fool significant numbers of "mainstream" computer users into attempting to load an NT-based product onto such decrepit, unstable computers as Packard Bells, Compaqs, and even ancient 486's! This will likely result in damaged equipment, deleted data, and the need to throw away a formerly useful computer and replace it with a "new and improved" machine running the slow, bloated Windows2000 code. By removing working computers from the installed base, Microsoft and their industry lackeys can squeeze out another round of unnecessary hardware purchases with yet another slower, buggier, more bloated and less-compatible OS.

Meanwhile, the efficient, reliable OS/2 Warp operating system just keeps going and going and going, like the Energizer Bunny of Software. OS/2 v4 provides more computing speed with far less hardware than the so-called "new and improved" versions of Windows, meaning less expense and less risk of data loss because there is no need to play the Version Game. Since OS/2 is still far superior to any Microsoft operating system in overall scalability, reliability, and compatibility with pre-existing applications and hardware, there is no need to succumb to the illusion of progress. IBM is focused on *real* progress, not merely the illusion. When you are the leader in the technology race, you don't need to try to pass anyone.... you're already ahead of the pack.

However, some people still ask, "Why doesn't IBM release a new OS/2 version?" Frankly, this question is really asking about marketing.... We might as well ask, "Why doesn't IBM change the name of OS/2 to something with a new year in it, the way Microsoft does?" The reason is quite simple: IBM is not in the marketing business, but rather the technology business. Renaming a product just for the sake of renaming it, to provide the illusion of progress, is a marketing trick. It does not provide additional value for customers, but rather the illusion of increased value. While IBM's customers know the difference, Microsoft's customers usually do not. The Microsoft method of marketing is to take a Yugo, add a spoiler and a new color of paint, and rave about "Yugo2000!" The IBM method is simply to provide a few extra horsepower, a little sturdier suspension, and keep selling the better product.

However, there is one very good reason to do a "new" or revised product release of OS/2 Warp. That reason is the huge amount of improvements, new utilities, updates, and third-party add-ons that have appeared over the past four years. Anything that would provide for a production refresh and bundle all of these enhancements into an improved, smoother installation would be a significant time-saver and provide real value for the computer user. With recent additions such as USB drivers, high-performance IDE drivers, improved Web tools, and APIs that focus on Linux and Win32 emulations, OS/2 continues to stay well ahead of the competition in terms of technological excellence. But the time and the complexity of managing the installation and the maintenance of these numerous patches is becoming a significant issue. Consolidation of these improvements, along with a more professionally-tailored installation procedure, would provide REAL VALUE for the OS/2 customer, both individual and corporate.

Lately there have been more "rumblings" about a third-party release of a "version"-type upgrade of OS/2. Whether this release would involve a numbering change (Warp 2000, anyone?) or merely a point increment (OS/2 4.1, perhaps?), there should be *something* different about the name of such a release, in order to demonstrate that a REAL VALUE is being provided by the update. My personal favorite is "OS/2 Warp 4.0 SE" or "Special Edition". This would avoid conflict with IBM's core customer base, who are gun-shy about major upgrades of something that is installed and operating smoothly. Yet it would allow sufficient version differentiation to show its improved value to potential buyers.

OS/2 continues to provide the finest computing experience on the Intel-compatible platform. This month's marketing noise and version games from Microsoft will not do anything to change that fact.

Most recent revision: January 30, 2000
Copyright © 2000, Tom Nadeau
All Rights Reserved.