In February 1951, Ferranti in England completed the first commercial computer. The next month, Remington Rand finished building its first Univac I. And by the end of the year, London food-service firm J Lyons and Co installed the first computer used for business data processing — running payroll, managing inventory and working commercial calculations.
And 50 years later, what do we have to show for it?
Well, we’ve got hardware and software — a gargantuan pile of hardware and software. Millions of tons of computers. Trillions of lines of code. And that’s just the stuff we’re still using today. The technology trash heap is even bigger.
More than that, though, we’ve got knowledge — a hard-won understanding of how IT really works.
For software, Brooks’ Law tells us that adding people to a late project makes it later. Linus’ Law says that adding enough debuggers to a project will speed it up dramatically.
On the hardware side, Moore’s Law says processors double in power every 18 to 24 months. And according to Metcalfe’s Law, networks grow in value by the square of the number of their users.
And where IT meets business, there are some other lessons we’ve learned, too.
- Automation can do only so much. When users’ brains control the machines, they’re immensely more powerful.
- Hardware keeps getting cheaper. Software keeps getting more expensive. The cheapest way to speed up the machine is with faster hardware. But the only way to make it more effective is with better software.
- Computing power isn’t measured in MIPS. It’s measured in dollars saved, sales gained and user effectiveness improved.
- Garbage in still produces garbage out.
- Even the best IT idea doesn’t stand a chance against corporate politics. But corporate politics are no match for the ingenuity of users. From PCs to the internet to wireless handhelds, users will drag business kicking and screaming into the future — which means users, not the IT shop, are the most powerful proponents of IT.
- We can build systems quickly or big or right — but not all three. If we build them quickly, they’re bound to be small or wrong. If we build them big, they’ll either be late or wrong. And for really big systems, they’re guaranteed to be late and wrong.
- No matter how much information matters to business, it’s still not what most businesses produce. Like the TV ad says, IT doesn’t make the airplanes and toys and food and cars and dishwashers — it just makes them better.
- Work expands to fill the bandwidth provided — and the storage provided, and the processor cycles provided. Yesterday’s "plenty enough" is today’s "barely sufficient" and tomorrow’s "hopelessly inadequate."
- Prediction is futile — and unavoidably necessary. We have to plan, but we’ll always get it wrong. That’s because technology may advance along a straight line on semilog graph paper, but what people will do with it goes in all directions — a zigzag, a tight curve, a widening spiral — and then jumps off the page.
- Users save us from our most catastrophic stupidities. They always have, despite bugs, delays and bad ideas. They compensate; they work around; they adjust, no matter how big a mess IT creates. After 50 years, they’re still the only ones who can.
Hayes, Computerworld US’ senior news columnist, has covered IT for more than 20 years. Send email to Frank Hayes.