|
Sun offers Java to ISOBut will Java make like Unix and split?
|
Mail this article to a friend |
Editor's note: SunWorld is published by Web Publishing
Inc., an IDG Communications company, independently of Sun
Microsystems Inc., which is not responsible for its contents. The
opinions expressed in SunWorld are those of the authors or
the publisher WPI/IDG and do not reflect opinions of Sun. The author is
not an employee of Sun, and none of the individuals quoted in this
article are Sun employees unless explicitly identified as
such.
Sun's official position on its PAS application is available
at
San Francisco (April 1, 1997) -- Half a year after Microsoft's high profile announcement of its plans to hand ActiveX over to the Open Group, Sun Microsystems has finally followed through on its pledge to make Java an open standard. On March 14, 1997, Sun transferred a formal application to the JTC 1 -- the joint technical committee of both the International Organization for Standardization (ISO) and the International Electrotechnical Commission (IEC) -- to begin turning over various unspecified parts of the language. JTC members will vote on the application, which is widely considered to be a shoo in, and the standardization of Java could begin as soon as July.
Sun is not saying exactly what parts of Java it will submit to JTC 1, claiming that it will "prepare a road map for submitting Java technologies" while its application is being considered. Two parts that will not be handled by ISO will be the brand -- Sun is holding on tight to that -- and certification. JavaSoft's is expected to announce tommorow that Utah's KeyLabs, Inc. will become the 100% Pure Java Program's certification administrator.
Sun's choice of the ISO as the standardization body for such a key Internet technology may also seem unlikely. After all, the organization that makes the standards on aircraft tires, widgets, and photocopiers isn't exactly known for its understanding of the Internet or for its lightning speed; and on the Internet, slowness equals obsolescence.
Mitchell says that the JTC 1 is a different story. According to him, the ISO set up the joint committee expressly "to speed things up." Another reason for choosing the ISO over, say the Internet Engineering Task Force (IETF), is because of the heavy weight ISO standards carry internationally. They are recognized on a formal basis by international governments, for example. Mitchell says that "no matter how we might have chosen to enter the standards process, the only sensible result is to end up with ISO as a standard." Eric Brown, senior analyst with Forrester Research, believes that ultimately, when compared to Microsoft's choice of the Open Group for its ActiveX standardization, the ISO "looks much better," because "there's a big difference in credibility between the two brands."
One thing that will speed up Java's standardization is the fact that the technology will be submitted to JTC 1 not via a working group, but from JavaSoft itself. In fact, Java would be the only technology to date submitted to the JTC 1 by a single vendor. So far, the JTC 1 has given its official Recognized Publicly Available Specification (PAS) submitter status to only four other entities: the European Workshop for Open Systems, X Open, the Digital Audio Video Council, and the Video Electronics Standards Association. The fact that Sun will not have to push Java through an industry association will undoubtedly make its adoption even speedier.
So not only will Java become standardized faster, it will also be more removed from the pesky and time-consuming matter of building consensus. While other interests may be able to veto the technology at the JTC 1 level and send it back to JavaSoft, it will be Sun and no one else who actually submits the various parts of Java to the committee. As Eric Brown says, "the stuff that's baked will go to ISO." To him, "this announcement doesn't really change that much. Java will continue to be driven by Sun and by JavaSoft at the pace that it's been."
|
|
|
|
Oh please, not another Unix
The trick for Sun will be to keep the user community happy by
providing other sources for Java technology, while at the same time
ensuring that the specification is not so loose as to turn Java into
"another Unix." Brown predicts that "the biggest problem will not be
defining pure Java, but defining the exceptions to Java where it has
to touch the real world."
Native methods could be one such area. Recently, for example, Microsoft was accused of forcing developers to use Windows-only extensions with its Java Software Development Kit. And Microsoft marketing manager, Cornelius Willis, accuses Sun of releasing a Java software development kit (SDK) that only works properly on Solaris. "One of the things that Sun has to do," he says, "is give us a version of (Java SDK) 1.1 that works on our virtual machines."
Mitchell is aware of the threat of fragmentation, which he calls the "open systems disease." He says that JavaSoft will prevent this from happening by, first of all, not losing touch with its industry partners and keeping them involved in the collaboration process at the API-development level. Secondly, he says that Sun will not let parts of the specification go into the standards process until they are mature enough. Of course, this reasoning assumes that industry partners, like Microsoft, will also want to play ball -- never an easy task.
ISVs may not be crazy about the move. Chris Tarr, vice president of research and development with Java ORB developer ObjectSpace, says that from his perspective, handing Java to the ISO is "probably not going to help me. It probably will hurt me." Tarr is concerned that if Sun elects to hand off a relatively immature part of Java, like the class libraries, it will slow down development of Java past the point that the market will bear resulting in, you guessed it, fragmentation. "The only way to stop something from fragmenting," says Tarr, "is to get there first. If the standards committees come in later then they fragment the market." Tarr says that this is what happened with Unix, and then, "it took a long time for Posix to come along and clean up the mess."
Whether Sun will make the right moves and prevent Java from becoming the next Unix remains to be seen. At least for now, it has diffused the embarrassing criticism that despite its constant harping on the importance of open standards, it was keeping this most visible technology to itself.
|
Resources
If you have technical problems with this magazine, contact webmaster@sunworld.com
URL: http://www.sunworld.com/swol-04-1997/swol-04-java.html
Last modified: