It's the Information, Stupid!

Esther Dyson

Where's the recovery going to come from?

To economists, it will come from macroeconomic factors: interest rates, investment cycles and the like. People will tire of being depressed, prices will become so competitive that it will be hard to resist buying and so on. ...

But for the IT sector, as for any other, there need to be specific reasons to spend money. Most chief information officers get a flat budget simply to keep their systems running. If they want any increases, they need projects to justify them.

And those kinds of increases will fuel the recovery.

How will they be spent?

We already had the long cycle of PC adoption. Call it self-empowerment. Then there was the Web: entertainment, media, the Internet as advertising medium. You can point to and eBay all you want, but this model was not very successful overall.

No, it's time to get back to basics. Technology needs to provide better information about the real world, and better ways of sharing that information across systems.

Over the last few decades, information systems people have developed a huge infrastructure of systems full of data about customers, finances, inventories, whatever. But most of these systems aren't well connected to the real world.


In this new age of security consciousness and interconnected systems, our information infrastructure needs to have a better understanding of the environment to keep track of it.

For example, the U.S. government's Sept. 18 draft report "National Strategy to Secure Cyberspace" was created to guide public action, but was woefully short on specifics. It did, however, point in the right direction, toward better information collection and sharing. Any canny CIO hunting for a budget increase now has a pretty good rationale.

Companies and governments need to know a lot. Who's using their systems? Where are those spare parts that were ordered? When are they expected to arrive? What if they don't arrive on time, if, say, they're stuck on the dock in California?

Today's systems need to deal with the real world in all its messiness, complexity, richness and sheer unpredictability. That messiness is not just from complex software; it's from complex reality.

If your system needs to talk to another system, you need a common language. For example, simple words such as "driver" or "renter" or "passenger" become important. How else can systems pass (with the user's approval) the identity of a frequent flier from an airline's Web site to a car rental Web site so the car can be ready when he changes his flight? For years, we've been reading about software interoperability; now we need data interoperability.


So what will CIOs be buying?

Not new systems such as those for accounting or inventory management. They'll be buying tools to help them to get more value out of their existing systems and to help interpret data that already exists but is currently beyond reach.

This may sound like a narrow focus. But the range of needs is broad enough to keep IT buyers and sellers occupied.

For instance, companies will buy catalogs and directories to define their inventories across vendors. If I'm short of part XG-3/8092 from MultiWidgets, what's the equivalent part from MacroWidgets? And if I can find a third option, from MonsterWidgets, maybe I can get MacroWidgets to lower its price.

I can tell where my parts are. But I really want to know (with appropriate permission) where my supplier's parts are. That's how I'll know if the finished product will get to me on time. Indeed, everyone from airlines worried about the reliability of parts to law enforcement worried about snipers is paying more attention to tracking goods and services around the world.

Here's another example. Hal Rosenbluth of Rosenbluth International, a global travel management company, says corporate clients want to be able to track their employees. If there's terrorist attack in Phoenix, say, they want to know which of their employees is in that city or heading there.

If there's a sudden need for a meeting, adds Jeff Katz of Orbitz (where I sit on the advisory board), clients want to know which location is cheapest, considering hotel rates and the sum of all airfares from each employee's location.

And all these systems need to be able to talk to one another not at the technical level, such as a software standard, but at a concrete level, such as a reference to a stock price.

There is a standard for systems to talk to each other called XML. But it's only the beginning, not the end. Juan's system can get stock quotes from Alice's system, but what does that "stock quote" mean? Is it the bid price or the asked? Is it current or delayed by 15 minutes?

This kind of cross-system data interoperability goes well beyond the to so-called "natural language" query tools of the past. They were supposed to let normal users find out almost anything simply by querying the system.

They failed to catch on broadly. It wasn't because the natural-language tools were inadequate (though they were). It was because the systems didn't have the information required. The data may have existed somewhere, but it was just not accessible.

Ten years ago, it was because the systems weren't connected. Now, they are probably connected - they're all on the Internet - but they can't communicate effectively.

Making all those systems more valuable - by making their data accessible across corporate boundaries - is the big IT task of the next few years. I predict it will lead the recovery. Many of the key players will be smaller companies acting as middlemen between existing large vendors and existing installed bases.

Larger vendors all want to sell multi-million-dollar projects. But smaller vendors are happy with sales of hundreds of thousands of dollars that simply realize the value in all the million-dollar systems already out there.


Страница сайта
Оригинал находится по адресу