OS/2 and PS/2: What might have been

James Cannavino spent Thursday just he does any other day of the week -- up to his ears in back-to-back meetings.

But when a caller reminded him that April 2 also marked the 11-year anniversary of the joint introduction of OS/2 and PS/2, he waved away suggestions that the date had special meaning.

"I never really think about it," said Cannavino, the chief executive of CyberSafe Corp., a Seattle supplier of security software for computer networks.

In April 1987, Cannavino was the key strategist at IBM's Personal Computer Co., as it pulled the curtain on one of the most ambitious product launches in its history.

PS/2 was the name IBM (IBM) chose for a new line of machines incorporating an advanced -- albeit, closed -- computing architecture. Sharing the stage was OS/2, an operating system that many experts then rated technically superior to both MS-DOS and the still-evolving Windows.

Have an opinion on this story? Add your comments to the bottom of this page.

Although IBM didn't mark Thursday's anniversary with any special commemorations, executives past and present talked about the event in ways that suggested a sense of loss, as if a major opportunity to write history had been fumbled away.

An expensive failure
The melancholy judgment of the last decade is clear: April 2, 1987 will be remembered as an expensive failure for IBM. But could it have been otherwise?

"It always struck me that had IBM, as an entire company, ever got behind OS/2, it [would have] had an opportunity," said one ex-IBM official, who asked to remain unidentified. "With the heft of IBM behind, that would have been something to see. The problem was that the hardware side of the company and the software side of the company stayed bifurcated."

To be sure, the balkanization of the company into different -- and sometimes conflicting -- fiefdoms didn't help matters. But Sam Albert, an industry analyst in Scarsdale, N.Y., and long-time IBM watcher, says the challenge had as much to do with the prevailing corporate culture at IBM as with anything else.

"Don't forget that IBM had a 'not-invented-here philosophy,' and that hurt them a lot -- especially when it came to embracing the technological innovations happening elsewhere in the industry," Albert said. "There are still a lot of people who think that OS/2 was the best operating system ever built, but it's not inconceivable that IBM spent hundreds of millions -- maybe even a billion dollars on all this."

For the record, Cannavino places the final tally at somewhere "between $1 billion and $2 billion."

IBM goes its own way
By the mid-1980s, IBM, the company that had legitimized the personal computer, was fighting off increasingly fierce attacks from upstart clone makers. The solution: IBM would go its own way with a proprietary hardware architecture called Micro Channel, or MCA. Although that move sharpened the divide between Big Blue and the competition, it failed to stem the erosion of the company's share of the PC market.

"We had looked [at the PS/2 launch] as kind of a watershed event in terms of the next generation of the personal computer," said Robert Carberry, who was the senior IBM official in charge of the project.

Now the CEO of CyberGuard Corp., a computer network security firm based in Ft. Lauderdale, Fla., Carberry says IBM's faith in the PS/2 was quickly rewarded. Within the first six months of its introduction, the line accounted for more revenues than the original PC had contributed to IBM's bottom line up to that time.

Some IBMers believe the company should have licensed Micro Channel to other computer makers, promoting it as an industry standard in much the same way Microsoft Corp. has set operating systems standards. But that was not IBM's way, and soon it had to contend with a new headache.

Competitors come out swinging
Nine other major PC makers -- led by Compaq Computer Corp. -- banded together with their own blueprint for advanced personal computing. Unlike PS/2, the so-called EISA boxes adhered to the Industry Standard Architecture at the time for PCs. What is more, the EISA standard also made room for slots that would accept add-on boards from third-party peripherals manufacturers.

That was where IBM went wrong, according to Carberry. Instead of presenting the third-party developer community with a take-it or leave-it ultimatum, IBM would have been wiser to gradually ease the industry over to its side.

"If we had included both kinds of slots, it would have provided for an easier migration path," Carberry said. "Things would have been more gradual and there would have been more co-existence . . . Strategically, our decision was correct, but a variation on tactics would have improved IBM's chances of success quite a bit."

Bad tactics also applied to the company's decision to price OS/2 at the top end of the spectrum, according to Stephen Manes, a technology columnist for the New York Times.

"In terms of OS/2, probably the worst strategy was the fact that it was the high-priced spread and wasn't very good in the early going and was competing against something that was cheap. That was the real nail in the coffin," Manes said. "You have to remember that IBM was charging a ton. If they had charged DOS-like prices -- or given it away -- they might have had a greater chance."

A more serious problem: OS/2 required 4MB of memory at a time when the average amount of memory on a PC was 1MB. What's more, memory was expensive and in short supply, and the average hard drive held about 30MB of data.

On top of everything else, the decision was made to launch the initial version of OS/2 as a product for 286-based machines.

"That really handicapped the product," Cannavino said. "It was clearly the unique duck in the pond and it wasn't a swan. I've done three full operating systems besides OS/2 -- and by the time I joined the project, it was so far down the path, it was hard to recover."

Time to bow out gracefully
One former IBMer linked the company's miscues to the mistaken belief it could dictate a change and the computer industry would follow right along to a new platform.

"They just didn't do the proper groundwork to get people to port," said the ex-IBMer, who asked to remain unidentified. "At the time, it looked as if it was IBM against the world, and they got isolated from the rest of the industry."

Despite ongoing promotions, PS/2 and OS/2 fell far behind competing products in the years that followed.

By 1994, IBM decided to merge development of its ValuePoint and PS/2 desktop PC brands, a move executives said would allow IBM PC Co. to use more common parts and designs. It was a graceful way to bow out.

Although OS/2 retained the fierce loyalty of millions of customers, that product, too, was put on the back burner by IBM.

It was finally time to close the chapter and move on.