Friday, June 25, 2004

Dvorak Predicted

This article appears in slightly different form in the May/June 2004 issue of IEEE Micro © 2004 IEEE.

This time I look at a ten year old book of predictions about the computer industry. Most of the predictions are wrong to some degree, yet the book provides a valuable look at a pivotal time in our industry.

Dvorak Predicts -- An Insider's Look at the Computer Industry by John C. Dvorak (Osborne McGraw Hill, Berkeley CA, 1994, 184pp, ISBN 0-07-881981-4, $16.95)

I met John Dvorak in the early 1980s. I don't think we've talked since then, but I've caught many of his radio and TV programs. I think we met at a reception put on by Media Alliance. In a room full of writers enjoying the late afternoon view of the San Francisco Bay and chatting, over drinks, about all manner of subjects, we were the only two who wrote about computers. It's amazing to compare that image with the jostling throngs of computer media representatives queued up to enter the hall for a Bill Gates or Steve Jobs keynote today.

In 1994, when this book appeared, Dvorak had been in the computer business as an entrepreneur, and then as a writer, for nearly twenty years. He had immersed himself in the computer industry. His columns dealt with nuts and bolts issues about computers, peripherals, storage media, firmware, operating systems, development tools, word processors, spreadsheets, and anything else you might think of. 

Dvorak's book is out of print now, but I found copies on the Internet for 80 cents and up. After all, who wants a ten year old book of predictions about the computer industry? The title is deceptive, though. Dvorak makes many predictions, but he also brings in a good deal of history and analysis to support them. Reading this book provides a fascinating look back at an important time in the history of our industry.

In 1994 practically nobody had heard of browsers and the web. The closest approximation to a website was a bulletin board system (BBS). The Java language had not yet sprung into prominence, and the closest approximation to "write once, run anywhere" was software or hardware emulation of the Intel 486 microprocessor. Online pornography was in its infancy, and junk email was a minor nuisance. Broadband to the home was a thing of the future, though a few people were using ISDN connections. Hardly anybody remained connected all the time. Most people dialed up their service providers when they wanted their mail. Even though Robert Morris's worm in 1988 had spawned the virus epidemic, malicious attachments were few and easily avoided. Rapidly spreading viruses and continually updated virus protection software were rare.

Dvorak does not mention Java, browsers, the World Wide Web, or spam. It's unfair to expect complete prescience, though. His analysis and some of his predictions capture the spirit, if not the details, of many developments that he did not complely foresee. He predicts a few things that have definitely not come to pass, but by and large, his predictions are good. In fact, the worst thing about the book is the many correct predictions he makes about things that do not matter at all ten years later. I think this points up the most problematic point about reading predictions -- it's rarely clear how to act to take advantage of them.

Though Dvorak didn't see the web coming, he saw something similar in his analysis of other people's predictions of 500 interactive TV channels. He said
It won't be 500 channels. It will be 50,000 or more channels all individually pumped out of homes and businesses in much the same way as computer bulletin  boards work today.

 . . . We can assume that people might just put a camera in their dining room, allowing us to watch a  family eat and argue.
He imagined that this might be implemented using BBS technology, with people dialing up over ISDN. From this he extrapolated to an amazingly prescient prediction.

Dvorak predicted that "Little brother will be watching, and little brother will be everywhere." From videotape of the Rodney King beating in 1992, to Amy Goodman's trickle-up journalism, to recent photographs of the coffins of war dead, we see how Big Brother cannot control the content of the news.

In 1994, Dvorak grudgingly conceded that we would have to take windowing seriously and that command line interfaces would die out. He felt that Unix, despite a nifty graphical user interface (motif), would nonetheless remain a niche operating system. He did not foresee Linux and the whole open source movement, though he did consider the possibility that the Public Windows Interface, as proposed by Sun Microsystems, might help keep Microsoft from building obstacles to competing products into its operating systems ("DOS isn't done until Lotus won't run").

Dvorak felt that Microsoft would try to achieve a monopoly in the industry. In 1994 it was not obvious that this could happen. IBM was still in the picture with OS/2. Apple was continuing to grow and was looking forward to a future based on the PowerPC architecture. WordPerfect was a popular competitor to Microsoft Word. Windows for Workgroups and its successors had not yet shown that they could push Novell from its network dominance. In fact, Dvorak believed that if Novell gave away DR DOS 6.0 to the entire industry, it could seriously undercut Microsoft. Novell did not do this, of course, so there is no way to know how effective that action would have been.

Dvorak also predicted that Microsoft would open a chain of software stores, so it could control both the shelf space and the sales pitch. This did not happen, of course, but that may be largely because Microsoft achieved those goals through other means. In the 1960s, IBM was amazingly successful at selling to data processing (DP) managers, which gave IBM market dominance at that time. IBM was a safe buy, so why risk anything else? Today's equivalent of the DP manager is the information technology (IT) manager, and Microsoft can reach them with a similar argument about networked Windows machines. On this basis Dvorak concluded that Macintosh sales to business would fall off drastically.

Dvorak's predictions of Microsoft's future were not all rosy. He said "Microsoft's domination will come to an abrupt end." He didn't say when, so he could still be right, but ten years later, there's no sign that this will happen. Dvorak's reasoning is interesting, though. He saw Microsoft as a company of youngsters willing to work 70 hour weeks and take gratuitous abuse from the management, because the rising stock price kept turning them into millionaires. He reasoned that the stock price could not keep going up without a few dips and that nobody would put up with Microsoft working conditions without that incentive. On top of that he says
Lotus has reemerged as a leader in the spreadsheet arena and pretty much owns the groupware category. WordPerfect is spending all its extra money marketing its word processor. . . . The best development tool are now produced by Borland and others. . . . NT does  not look like it will be much of a success either as  a network server operating system or as a standalone  operating system.
Dvorak also believed that hiring "old pros" would hurt Microsoft by eroding the general naivete and diverting energy into corporate empire building games. 

Dvorak devotes a large portion of his book to the microprocessor wars. He had great hopes for the PowerPC architecture. I was the guest editor for the Micro issue on PowerPC (Sept/Oct 1994), so I remember quite well the enthusiasm that everyone felt at the time. That architecture has been a success, and Apple has successfully moved to it. Dvorak believed that IBM had a secret strategy to include a clone of the Intel 486 chip as part of a PowerPC chip. I don't recall whether IBM actually did this. In any event, PowerPC has done little to affect Intel's dominance.

It is a cliché of futurism that people tend to overestimate the short term impact and undersestimate the long term impact of new technologies. In 1994 Dvorak said that voice recognition would be the killer app of the 1990s and would take off explosively in the next two years. This prediction is probably correct for the long run, but it hasn't happened as fast as Dvorak predicted. Similarly, Dvorak predicted that the Unicode character encoding scheme would lead to the death of ASCII by 1995. ASCII is alive and well today, though Unicode has become increasingly important.

Dvorak predicted that virtual reality would be a dud and that the TV and the computer will not merge. Dvorak saw that the marketing hype of the time overestimated the short term impact of these technologies. I suspect that Dvorak underestimated their long term impact.

I could go on and on. This short book contains a great deal of material. If you are at all interested in how we got where we are today, you should find a copy of this book and read it.