Age: 47
Claim to fame: Created the Java programming language.
What he's doing now: Fellow at Sun Microsystems Inc.
What has been the biggest technology influence on your life? I'd probably have to give you a really odd answer, which is drugs. If it wasn't for advances in modern medicine, I would not be alive today. In my particular case, about 20 odd years ago, I had a really bad case of amebic dysentery, and if it hadn't been for [an antibiotic], I would be dead. It's sort of a mega-antibiotic, and in many ways, it's been a much bigger influence on my life than any other thing, because I wouldn't have had a life.
What will be the next technology advance that makes a profound impact on the business landscape? I think the area of the most profound impact over the next three years, in some senses, is an obvious extension of what's going on today, namely the sort of deeper and more pervasive penetration of the network. I think we've only begun to see what can be done with the network, and in particular, the interesting ones are going to be the intersection of embedded processing in intelligent devices and wireless.
 |  |
| James Gosling, a fellow at Sun Microsystems Inc. | |
There's a lot of interest in this area in consumer electronics. You look at the back panel of most pieces of consumer electronics gear, there is a rat's nest of wires. And what travels on those wires, almost all of it is just information. You could get rid of the entire rat's nest of wiring just by networking these things all together through digital wireless.
But when, say, cars and phones are more deeply on the Net, when the positional stuff happens, then really the profound and radical changes are going to be sociological. And that's one of the key reasons why a lot of this is very difficult to predict because as a technologist, it's really easy to go through lists of bits of technology that could be put together. But then there's the social experiment.
What advance will change our day-to-day lives? One of the advances that could have a really important impact would be solving the last-mile problem. We're going through this great telecom bust right now of all these telecom companies that built in all of this huge backbone stuff that was based on optimistic projections of what traffic would be and what they could build for it. But, by and large, they sort of abandoned the last-mile problem because it was too hard. The closest thing to a solution to the last-mile problem has been DSL and cable modems, both of which feel to me like a Rube Goldberg solution.
You're hijacking one medium to do another thing, where the original medium wasn't really intended for that. It's really hard to get high bandwidth to your home. With cable modems, you can get moderately decent bandwidth but only so long as nobody else in your neighborhood is doing anything.
I know that there were a number of studies done some years ago into the economic feasibility of installing fiber directly out to businesses and homes. A lot of the stuff that I saw made it look pretty reasonable. So when I talked to people who were doing that kind of stuff, the problems weren't "Is there a business model?" or "Can the technology do it?" "Can we do it cost effectively?" Those things actually seemed to all be pretty well in hand. The issues seemed to be largely regulatory and who owned the right of way. The phone companies own the poles, and they don't want competition. The phone companies did a pretty good job of squashing all the little ISPs.
Maybe I'm just a technologist talking, but most technology problems actually seem easy to me compared to the social problems.
How will IT organizations change in the next 10 years? I would bet that in any fundamental sense they're exactly the same as they have been for the last 10,000 years -- big bureaucratic messes of people. I mean, the Egyptians had IT organizations. They just didn't have the information technology, like paper.
But that said, there are technology trends that are clearly pushing these org structures in different ways. Networks are a big piece of it. The fact that everybody is able to talk to everybody else, that the various infrastructure systems that everybody's building are much less isolated than they were. The presence of a network port is assumed on every box.
So IT organizations lose the ability to be separate and stand-alone. It feels like they have to merge a lot more into the actual divisions. People seem to be doing a lot less of their own software development, and I expect that to continue. Outsourcing of IT has been a mixed bag, but I expect that to continue pretty heavily. Outsourcing of e-mail has become relatively common.
What will be the dominant technology companies going forward? Well, I tend to have a bias, which ought to be fairly clear. I think that Sun is in excellent position. ... No matter what happens, Microsoft will do just fine. ... I think the software industry has broken up into a small number of big companies, and in some senses, only two -- Microsoft and Oracle. ... But almost all of the world's interesting software is done by the thousands of little boutique garage shops and ... little consulting outfits. As a community, I think they're going to win. Software is an area where you really can do a lot of your sales and business development on the Net. It's a business model that fits the Internet really, really well. And you find these little software companies in the oddest places doing fairly nicely.
And I really believe that the phone companies are going to do very well. They have to re-evaluate their position in the world and have a pretty significant internal restructuring and rethinking. But I think in the end, they will do well. I actually think that WorldCom is going to do really well.
Will there be any major changes in the development landscape going forward? I think the big change that's been happening is cross-platform integration, or end-to-end development, where people are not just developing for the box in front of them but [taking] this whole systems approach to development where you've got people developing software that goes in a telephone handset, that goes in a Web front end that's connected to that. It's just databases that are all working together to support that cell phone app.
And these things are a very different kind of distributed environment, and in terms of application developers' skill sets, that is pretty clearly the thing that is the most challenging for people to address. The developer tools for handling these things are just beginning to happen.
One of the things that I find somewhat depressing is that the economic model for tool companies is pretty dismal these days. You know, Sun got into tools not because we really want to get into tools but because as a software systems company, we are pretty dependent on them existing. We would be perfectly happy if people on the outside were doing really good tools. The problem is that they've all been failing. And some of that is because the market price for a really sophisticated tool has been set to next to zero, and because Microsoft uses their tools to control their developers, they make it extremely cheap, which means that nobody else can sell a tool for much more than Microsoft's tools.
One of the things that has always bugged me is your average software developer spends less on tools than they do on lattes. If you were to compare a software developer to a dentist, dentists spend a lot of money on tools. Developers spend none, or damn near none. And if developers spent more money on tools, there could be a really healthy marketplace for tool developers.
Can you predict which programming languages will be dominant during the next decade? Well, I think that Java is on tap for being pretty much dominant for a decade. I think in some ways it would be a tragedy if, by the end of the decade, Java was still the dominant language. I don't think that anything should live that long in this sort of area. I'd like to think that Sun and maybe me would have something to do with replacing it.
I think that the immense set of APIs [application programming interfaces] and tools that are in the Java world make it incredibly powerful and incredibly strong. And I think that whatever comes after will likely be an evolution on that, in the same sense that Java has been an evolution.
Do you foresee a splintered application development world between Java and Microsoft technologies? Yeah. It's certainly going to be splintered, and the splintering isn't really going to go away. I can't imagine Microsoft making peace with anybody.
Web services are the latest craze. Do you think they'll catch on, or do you think that whole movement could be a flop? The fundamental capabilities behind Web services have actually been around for a long time. One of the lines I tend to use a lot is: "Web services -- they were a good idea in the '70s; they're a good idea today."
Today's Web services are really CORBA [Common Object Request Broker Architecture] and RMI [Remote Method Invocation] and all these other technologies that have done basically fairly similar things only dressed in clothing of a different color. And one could certainly argue that the current Web services specs are way behind things like CORBA and, technologically, are a pretty strong step backward.
I think [Web services] are actually valuable enough that it won't be a flop. I don't think that doing distributed programming will be a flop. That's certainly on track to be a big success. It certainly has been for years. The style of doing it by shipping around XML fragments, I don't know. That certainly has huge performance problems that some older technologies don't have.
There are two big values in the whole Web services push. One is it's an awareness thing, getting people to think more clearly about building their systems not as isolated, self-contained universes. And the other one is that there's been this big issue over interoperability at higher levels of applications. You see this in all kinds of industries, and the sort of canonical area where people bitch and moan about it is in health services.
If you go to your doctor's office in the U.S., the nurse there will have a rack with 30 or 40 claim forms from all the different health insurance companies that service that area. They're all mostly the same but not actually the same. And if you look at the electronic analogues of them, they're not interoperable. There's no standard data type or external data representation for things like a blood report from the lab.
In the XML world, there have been a lot of people getting together to come up with a lot of these standardized representations for what is somebody's address or what is a purchase, what is a catalog entry, what is this, what is that. And you look at organizations like OASIS [the Organization for the Advancement of Structured Information Standards] that are collecting and organizing these standard schemas, I think that's going to be tremendously valuable, because then it will be possible for people to write applications that do things like talk to a doctor's office and schedule an appointment. And in some sense, I don't care whether it's XML, CORBA, smoke signals -- so long as there's a way to talk.
What kind of role will computers play in business 10 years from now? You can speculate over all kinds of stuff, and it's pretty clear that, in some sense, the track we're on is the obvious one -- more communications, more computers, people using it to grease the skids between things. A lot of the pressure in the IT industry has been around getting information delivered to the right people at the right time.
But what are the implications of perfect information? We're getting to a place where everybody will know what they need to know. And things will be able to react very quickly. If you look at the state of the economy and the state of the stock market these days, one could argue that one of the causes is this perfect information. Everybody's watching CNN or whatever, and they go, 'Oh, the stock market is screwed. I'd better sell my stock now.' So it turns into a stampede.
The system that we've got right now was never really set up or thought around what happens when everybody gets perfect information all the time immediately.
Do you think technology ultimately will be a net positive or a net negative? I think it will be a net positive, but I think we have to learn how to deal with it. There's a really interesting example of this [involving Global Positioning Systems in cars]. In Tokyo, they broadcast information about traffic loading on the road. They will color code every road as to the traffic density.
If you then had automated routing software that looked at the traffic density, the place would collapse because, if you're looking for the optimal way to go from here to there, and you know right now there's this one road that's got nobody on it, all of a sudden every car in all of Tokyo will converge on that one road and go down there. And it will just be a disaster. So they actually intentionally ignore information, or they'll have noise so that people will randomly choose not to do the optimal route.