As good as an operating system may be, it lives or dies according to its support, and support is something IBM is famous for. Yet OS/2 has always struggled to appear to be a well-supported operating system.

Although most of us know that OS/2 is alive, well, and going from strength to strength, it is very common to hear replies like, "...but OS/2 is dead, isn't it? IBM said so!" This coming from the competition is one thing, but from within IBM such information is very damaging. Most people don't appreciate the difference between companies within IBM: if the PC sellers within IBM kick OS/2 for fun or profit then that appears to be the "official" line.

The omission of OS/2 in important IBM plans for the future, or negative quips slipped into hardware launches may have appeared carelessness or stupidity on the part of underlings, had they not gone on and increased. We can only guess at the motives behind such tactics, but it is clear that they will continue, and that the old model of support for OS/2 -- the single, large company that is determined to stick with its loyal customers come what may -- is undermined. A better model would be based on multiple vendors, popularity rather than duty, and recognition of OS/2 as a foundation for value-added modules, not an end in itself.

There are aspects of the Linux style of support that are ideal for OS/2 now. IBM does some things well, but if you aren't one of IBM's major customers or don't have very standard hardware, then you have probably found that help from other users locally and via the Internet is the best solution when it comes to OS/2. Linux users are in a similar situation -- often vendors have no appreciation of anything other than their pet operating system, so if you have problems, the user community is a very valuable resource.

However, unlike Linux, the help you can get with OS/2 is limited by lack of freely distributable sources. So if a bug or new device requires software changes rather than configuration tips, you must rely on IBM to get around to it. Not that I am saying IBM is bad in this respect, but there are bugs from Warp 3, and ones found in the Merlin Beta, that didn't get fixed.

One question I have is whether some amalgam of present OS/2 and Linux styles of support would increase the chances that such problems are rectified quickly. The other question is whether the general public (and software authors) would view OS/2 as having better prospects if it was somehow "second-sourced" as far as development is concerned. I feel the answer is "yes" to both, provided the change can be well implemented.

Simply stopping development and handing over the code to some anonymous FTP site is not what I'm advocating. Something akin to development partnerships would be more like it, with some of the code made "copyleft" or distributable in a GNU-like way.

OS/2 is a foundation (now a very good foundation) for uses that must surely exceed IBM's imagination. I can imagine IBM keeping the respect of customers by continuing to provide OS/2 as a desktop system (even if the really interesting modules -- probably in Java -- are sourced from third party commercial or freeware authors). Having OS/2 as the bedrock in a file server or Web server or whatever also makes sense; if IBM hopes to develop software that must sit on an operating system it is strategically important that they not be at the mercy of a monopolistic operating system supplier.

IBM's OS/2 would then be like Caldera's Linux or RedHat Linux, sharing a common, well-supported foundation but still incorporating components and support that make it valuable to customers. IBM can still make money from sales of OS/2 as it does now; they can make more money out of add-ons as OS/2 penetrates markets otherwise unavailable; or they can simply use "integrated solutions" based on Warp Server to help sell IBM hardware.

Anything an IBMer might say against OS/2 then would carry less weight, given that IBM would then be just one development partner. The cost of the "base" OS/2 system would come down, increasing the market for add-ons. Cross-breading from Linux development efforts would mean more compatibility and early arrival of each new file system, etc. in the way that Xfree86 and many new applications benefit now. OS/2 would still earn IBM revenue, but now largely from the feature-laden versions, which they could even purge of the name "OS/2" if pressure within or without the company persisted.

IBM must be able to satisfy three requirements: maintain the image of a strong supplier that keeps its word, make a profit, and ensure it will stay in business. Presumably IBM understands the importance of its image, and OS/2 is said to bring in around a billion dollars of revenue. Possibly there are fears that even higher revenue earners are at risk by IBM's continued involvement with OS/2. Whatever the source of internal pressure on OS/2, an opening up of the product would meet all three goals, especially the third, and is infinitely preferable to running down the product in any way.

My own personal hope for the future is for the phrase "the operating system doesn't matter" to be taken for granted. We should be able to choose whatever is best for the job, then choose the user interface we want (text, voice, PM, fvwm95, etc.), and select from any applications out there. In some ways IBM's stated direction (using Java in particular) is in line with this. Making OS/2 widespread and at least the base freely distributable (e.g. enough to make a cheap network computer) fits in with such a philosophy.

But this isn't something to be entered into without careful thought given to the details. Don't take what I have said to be any sort of blueprint, but instead, consider it food for thought. Having OS/2 in exactly the same position as Linux wouldn't work, for example. This is also a step which I feel OS/2 simply wasn't ready for until now. Handled correctly though, this could benefit IBM and the computing community significantly.

--

[Our Sponsor: Post Road Mailer - A high performance, 32-bit, email program.]