While technologically sound, Jini is at a stage comparable to having fax machines but no phone lines.

"If anything, Jini is ahead of its time. We don't have the infrastructure in place yet...to really implement all this device communications stuff," said Patricia Seybold Group analyst Anne Thomas Manes. In particular, Jini is awaiting wireless networking technology such as Bluetooth and some crucial software ingredients from Sun itself.

Sun acknowledges that Jini has been held back by technologies like Bluetooth and a standard way of sending requests to a Jini printer. "It does take time to get everything aligned," Sasaki said.

Another sticking point: Gadgets don't have enough computing horsepower. Jini needs Sun's Java software to run, but it's not easy to squeeze Java into something smaller than a PC.

Unfortunately, Java for the small gadgets also doesn't have all the features necessary to run Jini. Sun's KVM version of Java lacks support for key Sun software called Remote Method Invocation (RMI). And moving Java to gadgets isn't easy. Sun planned to release KVM for the Palm in 1999, but it's still in beta.

Jini allows a way around these technical difficulties, though: More powerful devices can act on behalf of simpler ones. To that end, Sun has begun placing more emphasis on putting Jini on "gateway" computing devices from companies such as Cisco or Ericsson that also provide a high-speed conduit to the Internet.

Another way around the performance problem is to run Java instructions directly on a microchip instead of in a miniature computer that must first translate Java instructions. Though Sun's effort at building such a chip faltered, companies such as Dallas Semiconductor as well as start-ups Ajile Systems and Zucotto all are building chips to bring Java to small devices.

But Jini is more than technology: It is an illustration of how Sun operates, and that history has not been free of problems.

First, someone with an active imagination like Bill Joy or James Gosling comes up with a clever technology. Next, Sun figures out a way to use the idea to circumvent competitors instead of attack them head-on. Then Sun trumpets the idea as loudly as possible. Eventually, well after Sun has declared victory, the idea may catch on.

Take Java, for example--software from Sun that theoretically lets a program written once be recycled for use on numerous devices without having to be rewritten each time. "Java is great, but it's taken them five years to get to where we are today. Jini is going to be similar as a building-block technology," Hause said.

And if Java is anything to judge by, things could get prickly with business partners as Jini grows beyond prototypes and pilot projects. With Jini, Sun will be anxious to avoid the acrimony with companies such as Hewlett-Packard, Microsoft and IBM that has tarnished Java's history. The Jini partnerships likely will go more smoothly because the business partners aren't as likely to be direct competitors, as was the case with Java.

Java also has changed direction several times since its initiation in 1991 as a way to get electronic devices such as video game consoles and stereos to talk to each other. Java, inspired by the consumer market, now is perhaps most popular in back-end servers running e-commerce applications and is just getting started in sub-PC devices.

Jini is undergoing similar, though less drastic, changes.

Go to: The future: What Sun must do to rise