System/360's legacy

Computerworld - Forty years. Who'd have thought it would last so long? On April 7, 1964, IBM announced its System/360. The 360 wasn't the first mainframe, or the fastest, or even the most technically advanced. It was breathlessly billed by the press as IBM's bet-the-business gamble, a $5 billion investment that would make or break the company.

Actually, that last part was hype -- IBM already owned 70% of the data processing market on the day the 360 was announced.

What wasn't hype was that the 360 would change IT forever. And you're still living with those changes today. For good and for ill. Whether you're a mainframer or not.

The 360 idea seems obvious now: a line of computers that could all run the same software. But until 1964, that didn't exist. Every new computer came with a new architecture. To upgrade hardware, you had to rewrite your applications. And because the life span of a mainframe product line in those days was three to six years, that meant a lot of rewriting.

True, every time an application was rewritten, it got better. Bugs were fixed, designs were streamlined, new ideas were implemented. But there was no choice -- applications had to be rewritten every few years. That was expensive. And with a limited number of programmers in the world, that's about all they could do.

Enter the System/360. Suddenly, applications didn't have to be rewritten. They could be moved pretty much unchanged to new hardware. That saved programmer time and effort, which translated into saved money.

That was the obvious part. What happened next was more subtle.

Now that programmers had more time, they could do new things. And they could take on longer, more complex projects. Before the 360, a multiyear application development project would have been crazy -- by the time it was done, the hardware it was written for would be obsolete.

But hardware compatibility meant platform stability. That led to application longevity, which made complexity possible. A whole new world opened up for IT, a world of huge, business-changing megaprojects. For IT, and for the businesses we serve, that's been a very good thing.

The downside? In the pre-360 days, design decisions had a natural life span of three to six years before they were revisited during the rewrite. But with platform stability, those design decisions could last for decades.

That's how we got Y2k, isn't it? And why so many of our one-time megaprojects are now unwieldy, overly complex, hard to use and harder to maintain. They're full of design decisions that have outlived their usefulness -- but as long as the applications work, more or less, we can't afford to replace them. They'll live forever.

And in the 40 years since the 360's birth, 360-style stability has turned out to be such a powerful idea that after it mowed down IBM's mainframe competitors, it became the dominant paradigm for every IT architecture that followed. We've tried to re-create the 360 on everything from minicomputers to PCs and Macintoshes, from Unix to Windows to Linux. Today, we can't imagine IT without it.

That's the 360's legacy: 40 years' worth of stability, longevity and complexity that now permeate every platform we use in IT.

That goes along with 40 years' worth of obsolete design decisions and applications that can't be changed fast enough to truly serve the constantly changing needs of our businesses -- all because we're so stuck on that stability, longevity and complexity.

The 360's great strength has become our biggest weakness. And there's no sign we'll overcome it anytime soon.

So, mainframer or not, help yourself to a slice of the 360's birthday cake. The legacy of the System/360 isn't all good. But unless we find a better way to do IT, it'll be with us all for a long time to come.

Frank Hayes, Computerworld's senior news columnist, has covered IT for more than 20 years. Contact him at frank_hayes@computerworld.com.

Read more about Mainframes and Supercomputers in Computerworld's Mainframes and Supercomputers Topic Center.