by Esther Schindler

Should IBM’s OS/2 Be Open-Sourced?

Dec 07, 20078 mins
Data Center

A group of die-hard OS/2 users are petitioning IBM—again—to release the operating system’s source code as open-source. The question may not be whether IBM wants to do so… but if it can. Not, I expect, that IBM will actually say this out loud.

I have some skin in this game. Or, rather, I have a lot of scrapes and bruises, and a few scars. For some years, I bet my career on OS/2: writing books (savvy readers will recall my recent mention of Teach Yourself Rexx in 21 Days), magazine articles (there was a solid chance that any tech article you read about OS/2 had my byline), teaching classes about OS/2 to corporate users, and as an activist in OS/2 user groups. I’m a Certified OS/2 Engineer. Until I moved into my new house, three years ago, I had the world’s largest collection of native OS/2 applications—something over 300 of them, including a few that never saw the light of day. Yeah, I cared. I still care.

At a Comdex in the late 90s, I was in the press room chatting with three other computer industry journalists who had also written quite a bit about the OS. Jerry Pournelle wandered by. “Hey, Jerry,” said Wayne. “We were just talking about OS/2.” Jerry nodded and said, “Yeah, OS/2. Great operating system, with only one problem—”

And we all shouted in unison, “IBM!”

Technical Superiority

Everyone who tried it loved OS/2, especially OS/2 Warp. Technically, it was a wonderful operating system. It really multitasked. It was robust; yet, the OS would shoot itself before it would permit harm to come to a byte of your data. It had an attractive, truly object-oriented GUI (for its time; OS/2 2.0 screen shots make me wince, today). Sure, it had some technology weaknesses (I’m certain I can out-geek you on the details); but OS/2 was far superior to its competition. For me, at least, it was the first OS to achieve the Principle of Least Astonishment: when I didn’t know how to do something, my first guess was generally correct. (Mac OS X later exceeded in this principle; after several years of using Windows XP, I still spend five minutes looking for the simplest things.)

But you’ll note that I said technically. Between Microsoft’s business behavior (&deity had better give me extra credit for that diplomatic phrase), IBM’s inept marketing (which caused me to shout unprofessionally at perfectly nice IBMers who lacked the power to change things), and a computer press who’d been burned by OS/2 1.0 and believed Microsoft’s as-yet-unreleased Windows 95 would be far superior (because, after all, vaporware has no bugs) — well, technical superiority didn’t win.

OS/2 did get a new breath of life when IBM OEM’ed it to eComStation (eCS), which coincidentally has version 2.0 in beta. eCS is targeted primarily at enterprises that still rely on OS/2 for legacy software but need it to run on more modern equipment; yet, it still has a die-hard community of end-users who run eCS/OS/2 as their primary OS.

Get to the Open-Source Bit, Already

Anyhow. As I said, the committed OS/2 community sent a petition to IBM two years ago, with 11,613 signatures, asking the company to release the OS/2 source code (or whatever part IBM owns) under an open source license. “Sadly, IBM was ignorant enough to not answer our first letter, and this is why we sent a second letter to IBM,” wrote Kim Haverblad, founder of the OS2 World Foundation. They’ve added a bunch more signatures, and I guess are pounding on the desks a little louder.

Will it make a difference? If IBM didn’t respond then, I’m not certain it’s motivated to answer now. If CIOs who wistfully remember OS/2 and appreciate open-source happen to suggest the idea to their IBM reps, maybe the situation will be different. I wouldn’t mind.

If you didn’t use OS/2, and even if you did, you may be wondering, “Why bother?” How many developers and end-users would find ten-year-old innovations useful today? Obviously, more than 11,000 expect they would. The desktop user interface, called the WorkPlace Shell (WPS) is the most obvious example why.

But, Haverblad told me, Linux developers should also be interested, “Since what is really a sad history for Linux is the gfx GUI such as KDE and Gnome that really are crap compared to the OS/2 WPS.” The OS/2 code might be old, he says, but from an educational perspective it offers a lot of knowledge.

The problem is, I’m not sure that IBM has a good reason to release OS/2’s source code — or that it can.

First, OS/2 is a huge embarrassment to IBM, and large companies rarely cope well with embarrassment. They wanted OS/2 to “go away” long before they could contractually end support. Sure, they’ll release some code roughly associated with OS/2, such as ObjectRexx. Other code that was released to the Linux community was the code for JFS, from which the OS2/eCS community has gained a lot. But I think that most non-techies in IBM would be just as happy if nobody mentioned the subject, ever again. Remember: they never responded directly to the community the last time—only indirectly, such as in a Computerworld article

It also assumes that you can still find anybody who was involved and has the authority to make a “let it go free” decision. I’m still in touch with some of my old Team OS/2 buddies, but a lot of the innovative code was written more than a decade ago. I don’t think there’s a PR person assigned to OS/2 anymore (though I could probably personally tweak the dude who was responsible for it in the early 90s; Joe moved up to corporate PR). The original programmers are long gone; any IBM work done on the OS was moved to India some years back, and I presume that’s where it remains today.

Plus, the open-sourceable code would be messy and expensive to extract even if there were strong motivation. Real old-timers will recall that OS/2 1.0 was a joint effort between IBM and Microsoft. Who would want to revisit the old contracts to figure out what parts of OS/2 could be freed?

The deal was that IBM would market to enterprises, while Microsoft would sell OS/2 to individual users. Microsoft talked about Windows, instead. IBM did its part; OS/2 was adopted by a lot of financial firms, some of whom are using eCS today. IBM convinced companies like Lotus to develop for OS/2 first, which is why 1-2-3/G came out for OS/2 before a version was available for Windows. Excel just happened to be written for Windows. Yeah.

However, Microsoft is only part of the issue. During the heyday of Windows 95, Micrografx (which was acquired by Corel in 2001) worked with IBM to develop Mirrors—a software development kit that enabled ports of programs from the Win32 to OS/2 Warp platform. I’m not sure who owns that code—and no one else may be sure, either.

WordPerfect 5.2 for OS/2 used Mirrors as a temporary stopgap (though that kludge had some lovely WPS integration, unmatched by any OS features I’ve seen since) while WPCorp worked on WP 6.0 for OS/2. WP6 got dumped shortly before WPCorp was acquired by Novell (remember that? before Novell sold it to Corel?); I’ve always assumed that they dumped the OS/2 version to make the accounting books look better for the sale.

And then there’s the issue of whether they have the code or the contracts, which is even more embarrassing. (How would you like to be the IBMer who would have to admit, “Er, um, we lost it”? One ex-IBMer told me that one of the key bits of OS/2 source code was literally lost in the move from Boca Raton to Austin.) Tapes age. Managers and programmers move within IBM, leave or retire. Developers’ documentation and copies are destroyed or sent to sit unnoticed in storage facilities ominously remincient of the last scene in Raiders of the Lost Ark.

Haverblad says, “The primary reason for the petition is to get vital parts of OS2 open sourced to gain from the technology in current technology. If the gain then can be made in eCS, Linux or the future Voyager project that runs; well that doesn’t really matter; but it’s a shame if IBM lets the code go down the drain.

Do I want OS/2 to be open-sourced? Absolutely. Do I think it’s going to happen? Sadly, I don’t even think that IBM will repond to the users’ petition. But wouldn’t it be nice if they actually responded to the user community?—Esther, still a Team OS/2 member