losed door to competition. And while Compaq has developed more original technology than most PC companies, it hasn't developed a single open one that I can think of.
Nonetheles
s, even a broken clock is right twice a day. Mindful of this group's motives, let's look at the realities.
Why is the Internet bigger than Unix? Because you can depend on the Internet to be the same everywhere and you can't with Unix. Which way should Java evolve -- like Unix or the Internet?
I vote for the Internet. So should Java advocates. For Java to spread beyond a dedicated core, both the software industry and corporate IS will have to believe it has a stable future. Right now, Java has not realized its dream of "write once, run everywhere." Plenty of small differences in virtual-machine implementations prevent that. Sun executives talk quietly about revoking Microsoft's license if it ships its version of the Java classes. Who will be left out in the cold then? The few hundred thousand Java users or the hundred million Windows users?
In a standards body, Sun will not get everything it hopes for, and neither will Microsoft. That's the point: Companies will have to cooperate to grow the
pie. What do they have to lose? Let's examine the common myths:
Microsoft will hijack Java.
Sure, it'll try. For over a year, the rap from Redmond has been "Java is a language, not a platform." Sounds a lot like "The Internet is a network, not a platform," a rap Microsoft was forced to retreat from. However, should Microsoft not come to its senses, there's nothing about a standardized Java that prevents Sun or anyone else from putting out its version of correctness and letting customers decide. If this technology is capable of standing up to the rigors of mission-critical computing, it's capable of surviving a standards body.
Standards slow the development process.
Sure they do, in the same way that reading a map slows down the process of getting somewhere. It also prevents us from getting lost, a very likely outcome for Java if its advocates don't plan the route carefully. And what's the rush? Java is three years from being a mainstream application development platform. I'd like to
think it'll be right, not just available, by then.
Java's too young to strangle in a standards body.
Hmm, standards are too slow to achieve results, but too fast to allow growth.
Standards mute competition.
Really? Tell that to IBM, loser of the 1980s network standards race between Ethernet and Token Ring. Submitting its Token Ring technology to the IEEE 802 committee did not prevent the other main 802-governed technology -- Ethernet -- from whipping IBM in the market. But it did allow customers committed to Token Ring to know where the technology was headed.
No other operating system vendor has ever submitted its product to a standards body, so why should Sun?
No one else -- at least since UCSD's p-code implementation -- has claimed its product as a universal platform. Such a goal requires more than business as usual.
Right now Sun and Microsoft are mirroring each other, trying to establish de facto standards. I support the Java Lobby's call to Microsoft and Intel
to cease fractious activities. But I also think Java needs an independent standards body -- perhaps a combination of the Java Lobby and vendors -- dedicated to advancing the technology. If serious users are going to commit to Java, they're going to have to see maturity, stability, and reliability -- from the vendors as well as from the products.
Mark Schlack, Editor in Chief,
mschlack@bix.com
Raising the J
ava Standard
