October 5, 1998

Old Predictions

Making predictions is generally foolish since any prediction specific enough to be interesting will, invariably be wrong. Predictions that are the most influential are likely to be self-defeating, since preparing for a future changes it. Nonetheless, I'll give it a shot. In fact, if you look at some of the full essays and my Masters Thesis, I can claim a pretty good track record. My main failing is to be thinking too far ahead. Still, just a quick look at some of my older predictions. Rather than prescience, they seem to be the result of experiences I happened to have had.

• Online computing. I happened to have gotten into online services in 1966 at White-Weld. Dialing up a computer and interacting with it over arbitrary distances via a network seemed the norm.
• In the early 70's I started to explore federated systems. This is due to my experience in trying to extend timesharing models to multiple machines. It was a combination of my experience at Interactive Data as the built a computer center on the other coast and my Multics experience with shared memory. The basic idea was that read-only data could be "shared" efficiently by caching. This extended to caching data elements between systems that didn't trust each other. Distributed databases made naive an unnecessary assumptions which put much too high a value on perfect consistency.
• Personal computing. This is an aspect of federated systems since each was an agent of its users. I was used to having these systems to myself with no worry about the cost.
• Microtransactions, my Masters Thesis, were an obvious result of starting to think about charge for resources and realizing that users needed charges in terms of services they buy, not in terms of the raw materials. An obvious result of thinking about consumer services.
• Personal Computer II. In writing VisiCalc, we were creating a product and reveling in the freedom that came from not having an operating system reinterpreting the hardware. This was in stark contrast to the Unix world (including MIT's Athena project) stubbornly sticking to a model of computing limited by other's definition of what services an operating system should provide and their willingness to accept much lower performance.
• Electronic Mail. I did Lotus Express in 1986, a decade too soon. But I realized that we would have to sell the concept of electronic mail. Unfortunately, Lotus sold it as a terminal program instead. Lotus Express was also meant to be a basis for electronic transactions, not just person to person messaging.
• Networking. We are still in the early stages with the focus still being on the Web. But having been used to networking since the 60's, it is an obvious requirement. My effort to make home networking viable is still in the early stages.
• Answering machines, pagers, cellular phones etc. I don't claim that I invented them but when I first started using them they were shunned as geek or yuppie toys. I used them because of their utility despite this image. And, over time, they have transformed from toys for social outcasts to normal and even necessary tools.

What about my misses? Too many to discuss here. And, conveniently, I forget most of them.

Looking Ahead

OK, very briefly:

• Beyond PC's. The PC is in the late stage of its evolution. It is too well understand and defined. Rather than being being able to use the PC as a platform for experimentation, we must wait for Microsoft or others to deign to provide capabilities. PC's also suffer from ossification due to the increasingly complex interactions within the platforms and the applications . All the interactions between so many elements result in more ways to fail than to succeed. The successor to PC's will be cooperating devices. This next stage is necessary as we integrate computing into everything we do. The relationships between these devices must be "arms-length" in that they take responsibility for their own behavior even if the other devices fail. This resiliences is what has allowed the web to scale. It is a federation of systems rather than a single entity.
• Books. We've long talked about the paperless office, yet we keep generating more and more paper. As Xerox has noted, paper is a wonderful cache for online information. It liberates one from the constraints of the screen and provides for high resolution. But this will soon change as small and portable screens with high resolution become common. Also, between HTML and Microsoft Word (with its online viewing mode), the idea of formatting for reading on the screen is becoming common. There's still a need for more software to keep a trail of what is read but, unlike paper, the electronic medium gives us the opportunity to create new forms of organization. Electronic viewing has some major advantages over paper:
  • Portability. Having carried books on airplanes for reading and having read books on a small laptop (the Toshiba Libretto), the online form is much much lighter even including the computer. It also allows one to carry many books. The Palm Pilot is readily carried in one's pocket and can hold a number of books. We are soon going to see smaller high-resolution color screens on the smaller machines.
  • Online. It has become frustrating to read a document offline while isolated in the "tin can" known as an airplane. It is becoming a requirement that I be able to follow references or forward messages to others.
  • Enhanced Documents. We've come to accept books as static. But, online a book is dynamic and can be enriched with active content. I've become used to being to able to scan the contents and to find text by searching instead of relying on the index. The ability to provide alternative presentations is one that we have not begun to explore. Iillustrations can be interactive. Having data in a spreadsheet is much more useful than relying on the author to have to choose arbitrary presentations.
  • Updating. We are coming to expect that a technical document will have updates and that the examples will be readily available online. For now, these are complementary to the printed form but will soon be integrated with the online form.
  • Community. Amazon.com offers reviews of books by other readers and comments by the author. When appropriate, we will expect that part of the online experience is to be a member of that community of those interested in a given book or topic. The approach taken for best-sellers will be different from those with specialty audiences. But a lonely book will be a sorry sight (site?). It will also change the author's role. Some will be avid parts of their community and others will stay outside the fray as they release their children into the world.
  • Rearrangement. It is still necessary for the author to take narrative responsibility for a story. Having the user follow random links is not as effective for a story. Creating branches is often simply a gimmick. But there can still be asides, subplots and intertwined stories. But there will also be new forms more akin to serials in newspapers. So far, the online serial has been a flop. But that's more because reading online is dominated by the self-conscious act of being online. As the electronic form becomes more accessible, the barrier will change and we can then start really exploring the possibilities.
• Cellular Phones. It's fairly obvious that these are going to become the normal form of the telephone. From a purely technical standpoint, it is already viable to "wear" a personal phone at all times. What is changing is the pricing model. ATT led the way and others are following in making the cost of using a cellular phone competitive, if not better than the land phone. Israel provides an example of the effect of such a pricing model. As an aside, this is also a sneaky way to get rid of local unmeasuredservices (free local calls), but this is less of an issue now that we are starting to create a data network that doesn't rely on local calls. But competition will also bring back unmeasured services. In particular, there is no need for the traditional cordless phone once we stop charging for very local calls. The other barrier is that of protocols to allow for smoother management of connectivity and messaging. But these will come. The simple value of being reachable when one chooses, is overwhelming. Of course, this will subsume pagers and will become integrated with other messaging.
• Digital TV and other CE (Consumer Electronics) Futures. The current attempt to create Digital TV recalls other failures such as enhanced music CDs, Quad sound, AM Stereo, and so on. We also have Stereo TV for the few who notice. The problem is that we need to create a marketplace for innovating in Digital TV and for sorting out its relationship to the Internet. It will simply take too long to build a DTV infrastructure. The growth of the Internet will simply overwhelm it. Digital TV is simply a conspiracy to churn the TV hardware market abetted by a Congress that doesn't understand Moore's law – they have trouble believing that communications capacity is increasing very rapidly, they continue to manage it as a scarce resource (thus maintaining their power). They are still deathly afraid of laptops. Instead we need true digital connectivity that is not limited by the current entertainment infrastructure.