OS/2 has earned its freedom

IBM's problem child is no longer welcome at home. Others still value it, and should be given the chance to show why

"My client," said the defence lawyer, "comes from a broken home." It's not a plea for clemency that judges hear very often now that families fission like amoeba, but in OS/2's case it's still a strong argument. Conceived by Microsoft and IBM in the days when the two companies at least pretended to like each other, OS/2 soon found itself up against its half-sibling Windows NT in a bitter rivalry for attention.

There is no need to rehash the dramatic twists in the tale, except to note the tiresomely familiar sequence of commercial collaboration, accusations of betrayal, cash settlements and eventual demise. In this instance, Microsoft recently agreed to hand over $750m (£427m) to IBM in atonement for sins against the operating system and other Big Blue products. IBM is presumably using the cash to buy a large wreath.

There is also little need to dwell too long over IBM's own unforced errors in positioning, supporting, developing and marketing OS/2. For all its great potential — it really was "a better Windows than Windows" for a while — the operating system suffered badly from its single parent's vacillation over whether to push the PowerPC architecture as the basis for a desktop system.

Time and again the company had the right ideas: OS/2 Warp introduced integrated Internet support, Java and an object-oriented desktop interface, but none of them answered application writers' main question. "Why should we develop for you instead of Windows?"

Potential turned to frustration; frustration to abandonment. Getting on for ten years after the last significant update, IBM has tired of giving OS/2 houseroom. That's no reason to kill it; the decent act would be to make as much as possible open source — don’t worry, it won't compete with Linux — and hand it over for care in the community.

For despite IBM's efforts, that community still cares. OS/2 has OpenOffice and Firefox, developers’ conferences and commercial products. It has a time-honed solidity second to none, and no malware profile whatsoever. This is an important and valuable legacy, and IBM should honour what its problem child has become. Old software never dies, it merely opens up.