The Mobile Opportunity
Java's Rise on the Handset
Phipps: At JavaOne this year, every day we had a mobile device manufacturer or network operator launching a developer program—Nokia, Vodaphone, Sprint PCS. What I hear loud and clear is that the mobile industry needs Java software. They have 100 million handsets doing Java in the marketplace. It sounds as if wireless development might be working at last. But it seems that consumer applications and games are the big deal and that nobody at the table is equipped to profit from that market. How is it going to change the space we play in?
Renaud: Hats off to Sun for their leadership on this. Whether it's out of vision and a love for standards or a fear of Microsoft, God knows, it is happening and is a huge deal. Smartphones are where the action is. Microsoft has made inroads in PDAs with [Windows] CE, but that represents only a tenth of all the units out there with programming capabilities. So most clients these days are going to be Java.
Challenges still abound. Some of them, and perhaps the hardest ones, can't be solved within the Java community, but have to be addressed in a broader context that includes the carriers, the handset manufacturers, and, the Java community. But all these handsets have different form factors, different capabilities, different provisioning mechanisms, different networks, and different semantics for their messages. All that makes it very hard for the wireless-application world to get the same Web effect we were discussing earlier.
The fact that J2ME has made such inroads already gives us one base of stability on the client. That's the first step. The fact that we have a standardized application platform with J2EE, that's another huge step. Now, we need to harmonize the client form factors because that's a huge barrier to entry for application developers. We need to make sure there are incentives for interoperability at the carrier level. The Open Mobile Architecture (OMA), which was started by Nokia and brought together all the relevant players in the wireless space, is trying to do that. We have reason to hope it will start solving some of the hard problems. [Editor's note: Since the roundtable, the OMA merged with several other standards organizations into a new body called the Open Mobile Alliance.]
Kasper Brehmer: I think you're asking how we see the market developing services so we can actually create a business out of this. If we look at J2ME and all the formats and products at this point in time, we do have a fragmented market, and that's of some concern to us. That's also why Nokia is very much in favor of the OMA initiative. Most of the clients have MIDP [Mobile Information Device Profile] 1.0 as the base configuration but then make large or small extensions to it, which fragments the market. There will be differences in clients. There will be differences in user interfaces and the way smartphones operate. That's something we'll still see some years to come. I don't think you'll see the same standardized user interface on commercial consumer devices that you have, for instance, on PCs today.
One of the initiatives at Nokia is to minimize the number of user interfaces we have. We've announced that in the future we'll converge toward four user interfaces that target different market sectors. We also need to provide the tools to enable content developers to address these differences. Finally, we need to enable the service providers to be able to address the right market segments with the appropriate software. So, yes, we do see that there are challenges in the market, but they are challenges that can be addressed.
Seeking the Killer App
Phipps: In Japan, of NTT DoCoMo's nearly 6,000 service providers, the vast majority of them are providing games for the i-mode Java network. Does that mean that everyone at the table is excluded from the market for now?
Shelton: It's true that the handset market today is largely a consumer market, and consumers want things that will be entertaining. That is going to be a very important market, and we support it today. Borland is making money [on mobile development] today. We think that the standards work that is done to allow consumers to interact and play games over the network will require enormous server infrastructure that all the companies around the table are going to benefit from.
Benjamin and Simon talked about how successful Java has been in the mobile market. I'd like to take a counter position that Java has been completely unsuccessful in the mobile market because we haven't delivered to the mobile market a platform for Java on the device that does anything more than let them build one-button applications. MIDP is never going to let us build interesting applications on devices. We need to move far beyond MIDP before we're going to be writing interesting applications for mobile devices.
Walicki: It is not only games but also SMS [Short Message Service] that is taking off. What happens is that teenagers chat on the phone and they want to create a dynamic level of community as they move through the city. Web services come into play in the infrastructure, and it extends all the way from the heavy server back end. At HP, we see the intersection of rich media and mobile as what is needed in the future.
Pepper: We've got to treat the handset as part of the extended enterprise. It has to be device to database, and it shouldn't be treated any other way.
Phipps: If we scrub games out of our thinking, what will people do with these things in an enterprise context?
Renaud: They're already doing messaging. RIM and BlackBerry are extremely successful. Ultimately, the market will decide. The key point is that we've got to provide people with technology and a framework that will allow them to build applications. J2ME has been very successful in the sense that it's been adopted widely across the board, and that's a very real measure of success. The fact that we haven't seen a tidal wave of applications is due to other factors that have fractured the technology in the wireless space. Addressing that fracture has to be done outside of the Java community because it's not a Java-specific problem; it's a wireless-industry problem. The important point is we need to get the tools into the hands of customers so they can easily take their apps, bring them to wireless, and then we'll see what the users' values are.
McGovern: There are a lot of enterprise applications for wireless. FedEx has applications for its delivery people, but they're using proprietary technology. But will the CIO of FedEx want to buy proprietary hardware and build proprietary software for the next step? It's just going to be a matter of time. There's a huge demand for wireless, but the barrier has been very high. It is like when the laser came out in the 1960s, and it was a solution without a problem. I think we're looking of the same thing.
Mitu: Look at the automotive industry, with embedded devices in vehicles that let you monitor and diagnose a car remotely. With location-based services, we can know where a human being or a machine is. The technology is there, and there are many opportunities. We're not looking at how to play better games in this territory. Economically, there are solutions the market needs today, automotive being one of them.
Shelton: That's an example of the point that I was making about moving beyond MIDP. We have a working group of the JCP headed by IBM looking at how to create standards for Java in automotive telematics, but there's been no standard produced. As a community, we haven't been helping Java move forward in the device realm, and it's going to allow Microsoft to have an enormous entry to bring .Net and CE into the compact area.
A Return to the Client
Farrell: The applications we see today are ones that are necessary [for business operations]. FedEx has to have its mobile application. But what is interesting to the customers that I've talked to is the huge market for applications that are merely nice to have. Can I get a sales quote in real time over my phone? I don't necessarily need to do that, but wouldn't that be a cool addition to my applications. How do I design the applications so the same quote will be provided on an HTTP browser, a portlet engine, or a J2ME device. Now we get into a completely different discussion about how to architect the mythical one app that serves multiple different devices. I think that needs to coincide with this so that when the 2.5G and 3G networks happen and J2ME gets better, customers who want to do this will be able to.
McGovern: We're hearing two things. We've got one saying that the bases are all going to be different, that Microsoft isn't getting it because they're going to say everything has to be a certain way, which we're hearing is not possible. Then on the other side, we're saying that we want to write applications once to be deployed anywhere, so doesn't that lead toward more of a Java adoption than a .Net adoption in the mobile space?
Shelton: I don't think it's that Microsoft doesn't get it; I think they do get it. The problem for us, as a community, is that they're able to drive a set of standards very quickly and that they're going to deliver to the carriers a set of solutions for a mobile product that has a terrific development platform, has a terrific deployment platform, and really solves FedEx's problem. They'll do that much faster than we are moving as a community right now. The marketplace will not adopt the best technology solution; they'll adopt a technology solution that works. If we're not careful, that is going to be the Microsoft way in which everything looks like one device and they're all manufactured in the cheapest place and they all run Microsoft software.
Brehmer: I think you're discussing different issues. On the question of corporate applications, we believe that J2ME and MIDP Java is focusing a lot on the consumer markets, so that's where a lot of our focus is. However, I also think that the Java MIDP standard is an excellent tool for enterprise applications. We think that MIDP 1.0 actually goes a long way to kick-starting the market, but certainly there is lots of room for improvement. We're looking forward to the 2.0 version, which will alleviate and solve a lot of the issues for corporate types of applications, especially in the midlet area of the MIDP specification.
Benfield: When it comes to the PDA market, which company in the world knows how to sell operating systems for mass devices? Microsoft. Which company knows how to sell development environments for mass devices? Microsoft. I think the market reality is going to be that, as much as I'm a fan of Java and J2ME, Microsoft is going to clean up in the PDA market.
Phipps: But that's on the device itself. How do you get information to that device? What's the server infrastructure that's going to run that?
Renaud: All the handsets will never run in CE. That's just not going to happen. Talk to Nokia, Vodaphone, or anybody. They will tell you that we're in a very early position where Java is going to own the consumer market and Microsoft may end up owning the wireless enterprise piece—so there's sort of an inverted world. Who knows? The key point for us is to figure out the PDA story. If the CDC [Connected Device Configuration] profile doesn't satisfy the needs of enterprise customers like FedEx and UPS, then we have some very serious rethinking to do. If it does, then we stand a chance and we're going to go off and find out.
Phipps: I believe we're at the boundary of a new frontier of mass-market adoption of devices and platforms that target Java directly. It's back to the client, which opens up new markets for all of us.
Java has succeeded in large part through its community involvement. Each year when I've done something like this, it's struck me how easy it is for members
of the Java community to have an adult conversation with each other, and it thrills me to see that spirit of openness and cooperation even extends to what could otherwise be very partisan discussions. Thank you very much for the discussion today.
About the Moderator
Simon Phipps is Chief Technology Evangelist at Sun Microsystems.
|