W e b w o r d

IEEE INTERNET COMPUTING
Vol. 1, No. 3: MAY-JUNE 1997

HR

THE NEXT COMPUTER REVOLUTION

Charles Petrie

Stanford Center for Design Research
<petrie@stanford.edu>

HR

People are messy, especially when they are busy and productive.

HR

Some of us, at least, throw papers on any horizontal surface and forget detail and dates. That's why the value of computers has always been as an organizing prosthetic.

At first we thought the computer was just a fast calculator, and this image still fogs our vision of what computers could be. One thing they could be are bookkeepers for our lives.

COMPUTING BY NUMBERS

The application for the first electronic computers was computing ballistics. If we want to shoot a cannon over the hill, predict the weather, perform finite-element analysis on a bridge, or crack an encrypted message, a lot of numbers have to be ground finely and cleverly. Developing the algorithms for these important tasks is fine high science.

Bookkeeping is by contrast a rather boring job, but one that is fundamental to our lives. Everyone does bookkeeping: We balance our checkbooks. We keep track of the dinners we owe to friends (or they owe to us). We adjust stock portfolios in response to changing conditions. Commercially, we track our inventories. In academia, we balance the courses offered by our departments.

Bookkeeping means performing simple arithmetic in categories, and having a system that relates how changes in one category affect another. We count, we designate, we balance. Every day, hardly thinking about it.

It didn't take long for people to understand that counting and accounting, after number-crunching, were killer apps for computers. It was the US Census that got the commercial computer marketplace going in the 1950s. Univac, IBM, and others hit businesses with computers for accounts receivable and order processing. Computers can do the simple bookkeeping we're bad at: keeping track of everything, encoding our procedures, and executing them faster and more accurately than we could manually. Computers make complex systems possible, especially distributed ones.

BOOKKEEPING, NETWORK-STYLE

Distributed bookkeeping was the prize. It's one thing to have computer-generated bills and quite another to have an online credit card system, much less an airline reservations system. Try to imagine reconciling concurrent reservation requests without networked computers. You can't.

Now we have nationwide ATMs. Just-in-time inventories and virtual companies are based on connected systems keeping things in balance. And, at the higher system level, everything from "the Fed" to air-traffic control systems and thousands of important commercial databases rely on distributed bookkeeping systems to quickly count and adjust, almost invisibly.

Some of this is trivial and some of it is absolutely vital. We depend on a vast array of networked automatic bookkeepers. It has changed the way we live, for better or worse, much more than weather forecasts or bridge design.

Now that we're all connected via the Internet, we have all sorts of computer systems to help us manage our travel, finances, projects, companies . . . . Wait. No, we don't!

OK, so some Web services can remind us to send flowers on a birthday. So we can find airline schedules for a planned trip, or exchange information via corporate intranets. What we don't have is any notion of history, state, or change management as we move through our busy days.

On the industrial level, what we get as a result is a company like SAP linking up our legacy systems in ad hoc ways that require trained consultants to install and maintain. We get a few calendar systems that are just now beginning to be used decades after they became available. We get stand-alone PERT/CPM systems that do nothing to interconnect project participants as the plan changes and changes again. But this is nothing like what we need and could have.

COORDINATION REVOLUTION

Where is my pocket coordinator? The one telling me that Jack the architect just changed the balcony design, reminding me I ordered the steel for the original design and should check out the changes. The one telling me one of my suppliers added a component relevant to our new prototype. The one telling me of a change to the aircraft engine that lets me go back to a previously impossible fuselage design. The one that simply says I have to change my shipping procedure because of a change in company policy.

The alert reader will notice that these are all bookkeeping tasks. Track the inventory plans and the plane design; if one changes, make the proper adjustment and notify the affected person. Just like processing order entry. Bookkeeping, that's all it is.

The e-mail I use for managing my tasks is a long way from a system that would automate task management. Why isn't my calendar more automated? My travel planning? Why isn't my 14-year-old daughter's six-week project tacked on the Web so I can help manage her time? How about my own projects? If this is all bookkeeping, and computers are so valuable for bookkeepng, then why don't we have these systems?

Are we waiting for smarter machines? No, AI has been most successful exactly in helping with bookkeeping. Systems in use remember the expert rules and heuristics, generate good schedules, filter e-mail, and sort through Web-wide searches. None of this requires human intelligence, just tracking things. Are we waiting on better schemes for search and the management of dependencies? No, we've got scads of research on search and dependency management in design and planning, almost none of which has made it into the commercial marketplace. Almost nothing is new. I don't know what we're waiting on - do you?

Maybe we're still waiting for the same reasons T.J. Watson said "The world only needs three computers." It's easy to underestimate the power of simple ideas. Because the secondary and tertiary effects are hard to see. Because new systems are hard to get started. Because we're slow to change.

We're on the edge of the next computer revolution. Computers made modern airlines, banks, and government possible, all through better bookkeeping. Computer-based coordination of distributed projects is the Next Big Thing.

When it happens, the Internet will become a sea of services that understand each other and remember. Management hierarchies will further erode as peer workers are put in direct, coordinated communication. Agile manufacturing will be more than a buzzword. Business processes will re-engineer themselves. Massive new projects will radically decrease in cost and time. We'll all have more custom-designed lives, able to work with others more flexibly. And I'll know how and when to help my daughter with her homework project.

Just you wait. Better living through better bookkeeping is coming.

Also see

My interview with Tom Malone

Back to IC Online

Copyright © 1997 Institute of Electrical and Electronics Engineers, Inc., All rights reserved.