That's OK with his boss, Bill Gates, who wants the former Carnegie Mellon University computer science professor to concentrate on the future. As head of Microsoft's research organisation, Rashid oversees some 700 people working at five labs across three continents. From his vantage point, Rashid dismisses suggestions that personal technology has hit any kind of innovation wall. He envisions a future in which disk drive capacity of a terabyte is routine and user interfaces possess active intelligence.
What do you think will be the biggest changes in the way people use their computers today and say, five years from now?
I'm not certain it will be widespread in five years, but the biggest change in computer usage on the horizon will be the advent of "ubiquitous" computing. Just as the interconnecting of computers to each other through the Internet changed the way people used their computers, so will the interconnecting of smart devices in the home and office change the way we think about computing in our lives. In most major metropolitan areas, you should be able to have constant connectivity from your PC, laptop, tablet, PDA, etc., to the Internet and through the Internet to any of the devices that you own.
Will there be any difference in the relationship between devices and the people using them?
Devices will become more adaptive to their owner and their environment and more directly under their owner's control--no matter where the owner is. These devices would be able to communicate with each other and with your home/office PC and laptop and would have a common notion of who you are and what your preferences and needs are at different times.
Speaking of the future, Microsoft's initiative for an auto PC based on Windows never went very far. Was that a case of a concept being too early, or is it something that still could work?
The idea that when I get into my car it knows about me and my schedule, the traffic and how to get through the traffic, that notion is a viable notion. The question is, when will we get to the point where we have all the right pieces to make it work? I think it will take another couple of years.
What's behind your optimism?
The automakers would tell you that it's a case of iterative refinement. We already have 30 or 40 processors in the automobile that are connected on a fiber-optic network. It's clear that the car will have more centralised intelligence knowing more about you. It's just a question of getting to the point of picking the interface and the right forms of communication.
Devices are becoming increasingly intelligent. How do you think that trend will evolve?
Part of what I see happening is an integration of the user interface across all these different devices. Right now, people's lives are very compartmentalised as far as computers are concerned--there's me in my office, there's me in my home, and there's me in my car. These are all different versions of me as far as these devices are concerned. The job of all these computing devices is to really help users solve their problems and deal with the tasks they have and not to think of these as all separate experiences.
Can you provide an example how that would work?
This would be part of what we're calling "attentional UI," which means how to manage the user's attention and how to give the user better knowledge about what's happening. We're seeing it a bit in PCs where, for example, when a friend or colleague gets online and connects to instant messenger, you get a nonintrusive alert. Or with e-mail, where a small icon pops up when you have a new message.
But we're really just starting to use the real estate on the screen in a good way. In the highly connected, highly networked world, where there is a lot of information flowing, the tendency has been to have more event-driven kinds of interfaces. We've done some work in research here where you can put an events window on your desktop showing different events like weather, traffic and current news, and my schedule. This is always on my screen on the side of it, and when something changes, I'll see that and see what changes.
And what about in other devices?
People interface with a broader number of devices, such as cell phones and handhelds, and these devices have to figure out where something like an e-mail has to go. If I were in front of my screen at my computer, I'd rather see the message there. If I'm away from my screen, I'd like to see the message on my phone or handheld.
One of the things we've done is to effectively try and track what you are doing throughout the day so that your devices can recognise where to get the information to you. So you can take advantage of your calendar and your computer to try and build a model of what you are likely doing at any point in the day. We don't use that information now, but your computer should be able to tell if you are in front of it by keystrokes or mouse movements. Using that sort of simple location information, your computer can decide how and where to best present updates or alerts.
How would these devices be connected?
These are all distributed processing problems, and what we need is an infrastructure to help us overcome that-building an ecosystem of computing devices that can communicate, interact, and share basic information and can globally perform a task instead of thinking of them as individual islands.
Regarding the attentional UI, all this stuff about weather, news changing on my desktop already happens today. It's already there in my.yahoo. So, is the take-away here that the technology gets validation only if it's part of Windows?
No. The point of the attentional-UI discussion is this notion that information and notices can find you wherever you are in whatever circumstance. The fact that the underlying system knows something about you means it can get information to you in a form that makes sense in the circumstances.
So, it doesn't need to be Windows-specific? The underlying operating system could be based on Linux or something else?
Other people could do things like that. The concept is that increasingly, the compartmentalised nature of peoples' computing experience will be broken down.
What are your thoughts about how storage and hard drives are likely to develop?
We're probably two years away at most from a single hard drive with a terabyte capacity. We can have that now, but it's not necessarily cost-effective or elegant. That means that a person is close to being able to save every conversation or photo that they've ever had from the time they're born to the time they die on a terabyte of disk storage.
How much is multiterabyte storage on an average PC going to cost?
Basically, the cost of drives stays more or less the same in each generation, independent of the fact that they get bigger. I can walk into a CompUSA store and buy a 250GB drive for $299 after rebate. When the 160GB drives were first out, that's about what they cost. Most of the cost is in the electronics and the spindle. As the electronics and the technology improve, the cost of drives in each new generation will stay the same. When they finally introduce "tera-sized" drives, my guess is it should be about the same cost as high-end drives today.
And in the range of affordability for most computer buyers?
That's been the case, and I don't see any reason why it should change from 10 or 15 generations of previous computers.
Speech recognition, which got talked about a lot at Microsoft a couple of years ago, doesn't seem to get as much mention these days at the company. Will it get put back on the agenda or is it off on the horizon now?
We are moving ahead on it. Most of the focus is on improving the underlying technologies--and they are getting better.
How far off from the "Star Trek" scenario, where Scotty can start talking and the computer recognises him?
The "Star Trek" level simplicity was just a little bit on the odd side. Their computer could figure out when you were talking to it and when someone was talking to the others. That's a little far-fetched. Right now, we have a number of key elements in place, and speech recognition engines are getting very good at well-defined tasks. A lot depends on your belief about how people want to interact with computers.
For instance?
If you're a good typist, you can type faster than you can talk. And you are more precise when you type than when you speak. What you'll probably see is speech talking a role in settings where it's appropriate, such as in specialised situations where people are interested in using speech--such as with a tablet computer, where there is no keyboard. If someone's a good typist, there's no reason to use speech unless there's something wrong with their hands. In five to seven years, this will happen. But people will use the interface that makes sense for them.
|
This article first appeared as an instalment in the News.com Vision Series 3 special report. ZDNet Australia is a wholly-owned subsidiary of CNET Networks, which compiled the Vision Series for its flagship site, CNET News.com.
