HP works to reverse Sun Java victory

Sun gloated that Hewlett-Packard and others had been "stopped in their tracks" in January when a standards body voted against a plan to shepherd a standard for real-time Java without Sun's blessing.

But rivals are steaming ahead, despite the fact that Sun says the outside effort will infringe on its intellectual property rights. Not only will a specification be finished, but products conforming to the specification should ship by the end of the year, said Wendy Fong, standards manager at HP's embedded software operation.

Real-time Java would enable Sun's "write once, run anywhere" technology to be used in a variety of computer systems that have to respond instantly, such as factory robots. HP considers the technology critical to its strategy of using Java in embedded devices, but is dissatisfied with Sun's intellectual property rules and with the pace.

The two companies' differences illustrate Sun's difficulties in trying to propagate its Java technology. Another debate over controlling Java turned into a lawsuit between Sun and Microsoft, a dispute in which Sun currently holds the upper hand. Though it hasn't entered the courts, the disagreement between Sun and HP is no less ugly.

The depth of Sun's indignation over the non-Sun real-time Java effort is revealed in comments the company made on the plan during the voting process at the National Committee on Information Technology Standards, the standards body that HP and others approached to shepherd the standard.

"The proposed work cannot be completed successfully in NCITS because it will infringe on the intellectual property rights of Sun Microsystems Incorporated," Sun wrote in comments on the NCITS vote in January.

Sun also said NCITS approval "will not serve to further the spread and adoption of the Java Platform. It will have the opposite effect of splintering the platform into competing, incompatible versions and creating uncertainty about the future of this seminal technology."

Sun used strong language to persuade NCITS and its members against voting for the non-Sun effort. "We believe this proposal is a blatant attempt to forcibly standardize the property of one member of NCITS without that member's permission or cooperation...This is an extremely dangerous precedent which must be emphatically rejected by NCITS."

Apparently Sun and its ally, IBM, were persuasive. Lucent and Xerox were among those that changed their votes to "no" after an initial round of voting failed to approve the HP effort.

Despite the vote, though, there are still two separate efforts to write the real-time Java standard. Sun chose IBM to lead its effort, called the Real-Time Java Expert Group. The non-Sun effort is occurring in an ad hoc committee that's a subset of NCITS' real-time Java R1 committee.

HP, meanwhile, steadfastly refuses to participate in Sun's process--despite the fact that Sun offered an "initial participation fee" of only one dollar.

Its disgruntlement with Sun's process stems from the belief that companies contributing to the specification must turn their intellectual property rights over to Sun, Fong said. In addition, HP objects to delays, saying they'll have shipping products before Sun even has a specification. Sun has said that writing the real-time Java specification will require changes to Java, Fong said.

"Who knows when Java 3 is going to come out?" she asked.

What's next?
HP hopes that the independent effort to develop the real-time Java standard will result in a standard that is "consistent with" the standard that's coming out of the Sun process happening in parallel, Fong said. In other words, devices designed to comply with one standard also will comply with the other.

To accomplish this, there are several members of both the Sun and non-Sun effort who will try to keep the efforts aligned. In addition, the non-Sun effort has submitted its current draft specification to the Sun group.

Among the crossover members are Mitre, Aonix, and Rockwell Collins, Fong said.

Next week, the NCITS members who originally voted on the non-Sun real-time proposal will meet, and the ad hoc committee will hear feedback those members to hear about that voting. That feedback will be used to decide how to proceed at a future, as yet unscheduled R1 meeting, Fong said.

Meanwhile, also next week, the ad hoc committee will hold the next in a series of meetings to continue refining its real-time Java standard.