Five new technology trends to watch over the next few years:
1. Reliability. I know, I know, I keep on predicting that someday people will care about reliability in the computing industry. Had something really happened at Y2K, maybe that would have stimulated the demand. But, for whatever reason, people are not yet willing to pay more for reliable products. What's particularly sad is to see good quality manufacturers like DEC die off, while ship-it-when-it-compiles Microsoft thrives. But, there are techniques and technologies which can dramatically improve the reliability of software and hardware. In fact, increasing redundancy may be the only way we can take advantage of improvements in VLSI density -- look at Intel's hyperthreading CPUs, for example. The reliability revolution is at hand, we just need to demand it as consumers.
2. Magnetic Circuits. While most of humanity has been entranced by the rollout of the Internet, materials scientists have been making great strides towards room temperature (or close enough) semiconductors. But, aside from power transmission, not a lot of thought has gone into exploiting them once they arrive. Consider, for example, that the entire field of electronics, based on electric fields, might morph into one based on magnetic circuits. In particular, applications which can take advantage of the improved energy efficiency -- portable equipment, RF Hi-Q tanks, etc -- will probably evolve in that direction.
3. Wireless Last Mile. Something I've been predicting since 1987. There's a huge glut of bandwidth out there, but most of us can't get to it, thanks to the stalled Broadband Revolution. While DSL works just fine on new copper, we've discovered that the majority of local loops are either poorly documented or held together with bailing wire. DOCSIS isn't bad, and the cable companies are trying harder (they have to in order to compete with direct broadcast satellite), but it's only available in very urban areas -- and PPPOE sucks. So, a third alternative is needed, and IEEE 802.16 will hopefully provide it. Even better would be an Ultra Wideband mesh network, if they can ever get UWB off the ground.
4. Remote Presence / Realtime. Together with a new generation of microactuators and sensors, the wireless last mile will pretty much enable you to have a presence anywhere the network reaches. We're already seeing this for high value equipment such as locomotives and aircraft engines, but just about everyone could benefit from better knowledge of a remote situation. First, however, we need to replace the creaky old software and protocol tools we have today -- H.323 was good ten years ago, but there are much better codecs and underlaying protocols. SIP is definitely a step in the right direction. And, we need to introduce a little more intelligence to the router network, in an Open Standards fashion. Reservation and realtime protocols exist, but none of today's protocols are able to reflect the economic costs of resource reservation, so they can be hacked or at least hogged. Applying some of the microcommerce technologies here would resolve the problem.
5. Supranets. Supranets are networks of entities built on top of other networks. For example, one can deploy a network of sensors, or a network of interconnectable services, on top of the communications network. And, on top of such a sensor or services network, one can find networks of people, of which the Slashdot community is a good example. Supranets can solve a lot of problems, but not without some enabling technologies. For example, how do I find other nodes to network with? We already have such an intermediary between web authors and web readers -- search engines -- but need to expand these to new services and also to people. Another necessary enabler is a medium of exchange; microcommerce technology, electronic futures markets, even the Slashdot karma system are examples of what might work.