Java in Wonderland

History repeats itself -- the first time as tragedy, the second time as farce. What comes after farce? Read any article on the Java marketplace and find out

You read the articles. You follow the court cases. One day, Bill Gates is on cable television's public affairs channel (C-SPAN) defending Microsoft in front of a hostile U.S. Senate panel. The next day he is in Fortune magazine in a full-page advertisement for Callaway golf clubs (he finds the game "humbling," in case you were wondering). A few weeks later he is introducing the much-feared-by-the-Justice-Department Windows 98 at COMDEX with a demonstration that terminates with the dreaded blue screen of death. This halts the machine and initiates gales of laughter from a crowd that has seen the same screen once too often on their own computers. Everyone seems to be talking about this one. Even the Microserfs must be chuckling softly. Well, all except the one who probably got fired. By mid-May, Microsoft bowed to incessant government pressure, and in a stunning announcement, agreed to delay the release of Windows 98 -- at least for a day.

Judge Robert Bork, one of America's foremost legal scholars, former solicitor general, the man who fired Archibald Cox during Watergate (and who has been paying for it ever since), is now fronting a high-tech lobbyist group. The Project to Promote Competition and Innovation in the Digital Age (or ProComp) is an educational and political action group funded by Netscape, Oracle, Sun, and Sybase, among others. Also advising ProComp is the former presidential candidate, honored war veteran, and otherwise unemployed husband of the much-loved Elizabeth Dole -- Sen. Robert Dole. Oddly enough, this group seems to have an interest in Microsoft's antitrust case. Just as suprisingly, some of the most conservative minds in America are suddenly lobbying the federal government to rein in one of the free market's biggest success stories. Curiouser and curiouser. Even Bob Dole has to admit "We're not in Kansas, anymore."

Meanwhile, in JavaLand, Hewlett-Packard announces that it is supporting 100 percent pure Java -- except in computers, embedded systems, and pretty much anything else with a battery or power cord. Then Sun states that it is working actively with HP to develop a common standard, although HP retains the right to go its own way if it wishes to do so. A few days later, Sun introduces the Java Plug-In, which will promote consistency across virtual machines; any virtual machine, especially the one from Apple, which is collaborating with Microsoft in Java technology. Apple, it would appear, supports four virtual machines at the moment: its own, one from Microsoft, one from Netscape, and one from Roaster Technology (which recently has been experiencing its own difficulties).

A little further down Route 280, an appellate court in Santa Clara rules in favor of a Sun injunction to stop Microsoft from using any Java compatibility logo. This is a major victory for Sun. Of course Microsoft's Visual J++ is already number one in the Java market. No doubt, several hundred miles north in Redmond, an able lieutenant approaches Chairman Gates with a smart salute and the report "Mission accomplished, sir." Just to add to the confusion, Sun reorganizes JavaSoft into two separate divisions.

Okay, let's face it, Alice. We're down the rabbit hole. Behind the looking glass. Only Lewis Carroll could be writing this story -- with the possible exception of Jefferson Airplane, of course. Can you hear Grace Slick in the background? One pill makes you larger. One pill makes you small. And the one the market gives you leaves Java no place at all.

What's going on here?

First: The Java Plug-In is at best a short-term solution. It may provide compatibility across virtual machines, but that is what the Java VM was supposed to do anyway. Perhaps most importantly, the Java Plug-In is just that: a plug-in. It requires each environment (that is, browser) to be retrofitted with an additional piece of software. Does this remind you of client/server technology? Internet technology has been successful in the corporate world in large part because it gets IT departments out of the software distribution business. They can deploy applications immediately to all users without requiring a specific runtime environment on each target machine. Plug-ins, like client/server runtimes, defeat this purpose.

Second: Hewlett-Packard is executing a Switzerland strategy with its own Java initiatives. It has very successful product lines that use Unix, Windows, and embedded operating systems. It will use the best technology for each platform and thus continue to be a true agnostic. That is what customers want: someone to make the hard decisions for them. When and if the Java marketplace settles down, HP will side with the winner. In the meantime, it will do what seems to be right for each unique customer base. Lots of corporate IT departments will consider it a very good strategy.

Third: The resolution to the crucial aspects of Microsoft's legal cases will lead to a more vibrant marketplace. A lot of risk captial in Las Vegas is being placed on the break-up of Microsoft into separate companies. This won't happen today. It may take a long time to evolve, but ultimately it could take place. The result will be a big win for all concerned, especially Microsoft stockholders. The restructuring of AT&T and the Regional Bell Operating Companies (RBOCs) provides an excellent example of this. The so-called break-up of AT&T spawned a number of top-notch companies as well as the world's most vibrant communications marketplace. This kind of situation certainly would not be a loss for Microsoft; it could be a win for all.

Fourth: Short-term success for Java is increasingly dependent on the server, not on the client. This has been apparent for some time. The server space is where the real applications get built, and where the real power lies, and Java is making significant inroads here. Performance is still an issue, but this gets better as each month passes. Java is an excellent language. Put the political maneuvering aside and this fact remains. The corporate IT community is becoming increasingly committed to component- based, object-based solutions -- and Java is the best language in the world for that paradigm.

Fifth: A unified Java will emerge in the long-term. This is small consolation to the besieged Java developer today, but it will happen. It is what the market wants, and what the market wants it gets. This is not to say that the path to a unified Java won't continue to be difficult. There will be many sharp turns and dead-ends. But we will get there.

What's coming for Java?

In short, even behind the looking glass, Java appears stronger than ever. As we move into the second half of 1998, there are really only two limiting factors in the ascent of Java as the world's most popular programming environment.

The first and most serious is true cross-platform execution. Even assuming that a unified standard does not emerge over the next two or three years, Java already is the most portable environment around. This is true even if the market, legal system, or political system doesn't create a single standard.

The last remaining obstacle to Java's ascendance is performance. That situation is getting better on a daily basis. Major initiatives are underway to close the performance gap with the industry's most reasonable reference language, C++. Significant performance obstacles like the JDK 1.0.2 broadcast event model have been streamlined with "listener" objects. Garbage collection is being optimized with Sun's HotSpot technology. This approach utilizes generational sweep algorithms to manage the heap based on grouping objects by age (and assuming that a "nursery" of newly created objects should be swept more often). Compilers, whether they're of the source code, JIT, adaptive/dynamic, or bytecode optimizer variety, continue to improve. Many hardware companies (eight at last count, not including Sun) are developing Java chips that are expected to be priced at less than 0 each. This generation of microprocessors will greatly improve performance as well as create a hardware version of a standard VM. The list goes on and on. In short, the performance gap is closing. Java doesn't need to be the fastest language on the planet, just competitive. If you think it hasn't reached that state yet, check back at the end of this year.


In the end, developers will write fast and portable Java code, Judge Bork will go back to writing legal tracts, Bob Dole will work on his wife's future political career, Hewlett-Packard will prosper, and Bill Gates will be smiling. Like the Cheshire Cat, even if Microsoft as we know it dissolves into thin air, Gates will have much to smile about. A new group of independent Microsoft Operating Companies (MOCs) will create a larger, more vibrant market, and some of them will be producing Java software. Java will create countless new millionaires -- and a lot more buyers of Callaway golf clubs.

William Blundon is executive vice president and co-founder of The Extraprise Group (, a leading provider of application development, training, and strategic advisory services for e-business. His focus in the last nine years has been on distributed object environments and the Internet. He is a former director of the Object Management Group.
1 2 Page 1
Page 1 of 2