Opinion: Java-based wireless carriers try too hard

To see what's "wrong with wireless" you need look no further than the just-concluded JavaOne show.

All the vendors sounded like cheerleaders on the sidelines, urging onward the team of enterprise users on the field, facing three-time defending champion SUD - skepticism, uncertainty and doubt.

A panel of vendors detailed alleged "challenges," like the need for high quality graphics, more dynamic applications, richer content, lower cost, higher performance and longer battery life. Some of the vendors said that their next generation of keen microprocessors would overcome lots of these challenges.

Sun Microsystems Inc. announced two applications that it promised were a "breakthrough in consumer device technology" - mobile phones and cheap PDAs. What could these marvels be? Not one but two, yes two, new Java Virtual Machines (JVM).

Codenamed Project Monty, the two Virtual Machines, in the wonderful marketing-speak that was abundantly present at JavaOne, "are expected to deliver a level of performance up to 10 times higher" than the other JVMs available now.

Sun vice president of Java and XML, Richard Green, twirling his own pompom, said a U.K. market research company, ARC Advisory Group Inc., had concluded: "By 2004, there will be over a billion Java-powered devices in the world." I'm not sure where that figure came from. The ARC Group Web site has a press release about its November 2001 report, which predicts 442 million Java entertainment users globally in 2004, and 1.1 billion mobile data users in 2006. The report refers to "Java-enabled" devices, not "Java-powered" devices.

The distinction is important: it's the distinction between "infrared-enabled laptops" and "infrared-powered laptops." The issue isn't how many devices have this stuff: it's how many of the devices can actually make use of it, in a way that brings an identifiable value to the enterprise.

"This is the year for wireless Java," Green was quoted as saying. One would think after the litanies of "This is the year of the LAN" or "of client-server" or whatever, Green would be a bit more cautious, if not more realistic.

Carriers like Sprint PCS Group were part of the cheerleaders, too. A couple of executives waved megaphones and pompoms and said the company plans to create the "phone you use for a lot more than phone calls."

This is not just an experiment," Sprint Senior Vice President Scott Relf and Vice President John Yuzdepski said. "It's a wireless revolution." The "R" word.

To whip up JavaOne attendees into a handheld development frenzy, the conference featured a "cutting-edge competition" built around the Sharp Corp. Zaurus SL-5000D, a high-end PDA running Linux and the PersonalJava libraries and APIs. Contestants had to "collaborate" to find bits of Java code. The winner received a varsity jacket with the Java logo.

A Hackathon let developers create their own Java handheld application for several categories: best user interface design, best design, best productivity application, most visionary application and my personal fave, "most useless but really cool anyway application."

The best productivity application was something dubbed the Conference Navigation System, which created a Zaurus-based mapping system for the JavaOne conference and Moscone Center, the show's San Francisco venue. The user specifies his or her existing location and a desired destination. The software creates a map showing the best route.

There was no indication whether the map that was created was, in fact, the best route. You can find these other notable programming achievements here.

Sun also introduced the "authoritative, central resource dedicated to helping Java technology developers successfully develop, deploy and make money from wireless Java applications." In other words, yet another Web site with some bulletin boards and tutorials.

To show "yet another example of [a] real-time Java end-to-end application," James (The Father of All Java) Gosling participated in an onstage battle between two robots guided by commands issue via Java-based cell phone keypads.

In the midst of all this hoopla, one Java user, Mike Walker, director of worldwide research and development for the U.K.'s Vodafone Group PLC, gave a presentation on Java and wireless security. Walker is also chairman of a security committee of the 3rd Generation Partnership Project, a wireless industry group.

"This sandbox approach [the touted hallmark of Java's security model, which restricts executing code to a quarantined area] is no security at all," Walker told his audience. "It's a joke. So just forget it." That's the security approach used in a subset of Java 2 Micro Edition, called the Mobile Internet Device Profile, or MIDP. A Sun marketing manager later said MIDP was "perfectly adequate" for the types of devices in use today.

Then a reporter, Matt Berger, with Network World's sister wire service, the IDG News Service, sidestepped the cheerleaders, the coaches, and the front office suits, and talked to the players on the field, the enterprise developers.

"There's been a lot of talk about wireless here," said one. "Wireless may be in the future, but right now it's just hype."

Another software consultant, who works with banks, said, "Some of our customers have an interest in wireless, but I can't say the level of security is ready."

"We're looking to see how we can use wireless in the enterprise," said another.

There's a disconnect here, in my opinion, between the "revolution" the vendors, futurologist and other pundits see, and the workaday enterprise issues like security, performance, data integrity and all the rest of the details that network executives have to deal with in deciding how, and whether, wireless Java fits into their business.

Sun does enterprise users a disservice when it encourages a "goldstrike" mentality among Java developers. Carriers are trying to find a magic formula for boosting data traffic to pay for costly spectrum licenses and 3G network infrastructures. Those aren't the problems that network executives are trying to solve.

And they're not going to solve those problems with robot battles, and Java varsity jackets.

Join the newsletter!

Or
Error: Please check your email address.

More about ARC Advisory GroupSprintSprint PCSVodafone

Show Comments