Sun's launch of its Javastation NC last month, billed in the press invitations as the "announcement of the decade", was an exuberant affair. With almost evangelistic fervour, Sun unveiled the products and technologies which it believes will usher in a brave new third age of computing.
The reason for Sun's passion goes beyond the excitement of simply participating in a revolution. The company perceives itself to be in the vanguard of this new movement, and looks forward with anticipation to reaping the rewards which that status will bring it.
Throwing common decency and restraint to the wind, it took the opportunity to laugh in the face of those it believes it will supercede. In an updated parody of an ad for the Macintosh that Apple ran in the mid 1980s, Bill Gates was vilified as the dictator of the desktop, a complacent Ceaucescu who was about to find the masses revolting against him as he looked on in bewilderment.
But as CEO Scott McNealy began to explain what he described as the third great paradigm shift of the computer age, it suddenly struck me what was wrong with Sun's message. The logic that we are entering a new phase in computing is irrefutable. But, in another parallel with Apple in the eighties that wasn't highlighted, Sun's claims to be at its forefront have got it wrong. The company has read the right runes, but its products are not the ones that will benefit from the new computing landscape. That fortune will fall elsewhere.
Enabling forces
It was when McNealy spoke of the three conditions that were coming together to enable the new paradigm that I realised which part of his argument didn't stack up.
The first enabling force he cited was bandwidth. I had no quibble with that one. People may complain about the slowness of Web access and downloads, but if you step back and take a slightly longer-term view than the current few months, it's obvious that bandwidth is coming onstream in a big way.
Just a few years ago, 9.6k was the top-of-the-range modem connection on a PC. Then 14.4k became the ultimate for a year or two. Last year, that became 28.8k, now it's notched up to 33.6k, and 64k is already on the horizon. This is without the benefit of upcoming technologies like ADSL, and still running on lowly copper analogue lines, never mind the potential offered by ISDN, cable or fibre optic.
In the network and leased line environment, developments like fast ethernet and ATM promise to open the sluice gates within a year or two. Ignore the naysayers that tell you the internet backbone is about to collapse. The bandwidth capacity of the global telecoms network is virtually infinite. It will quickly come on stream for data communications as demand rises.
The second factor was the Web. This assertion I was equally happy to accept. In less than a year, the IT world has fallen head over heels in love with this beguilingly simple method of delivering information services across the enterprise.
Java solecism
So far, so good. McNealy was right to pick out these two as significant and revolutionary developments. Then he came to the third. "Java" he proudly declared.
This was such a jarring solecism that it stood out like a sore thumb. Java is not a trend or even a technology. It's a product. Whatever its merits, it's simply going too far to hang a whole paradigm on it.
Of course, it would be uncharitable to chastise McNealy for this very understandable error. Java is his company's baby, and as the proud father-in-chief, he can be forgiven for allowing his paternal enthusiasm to get the better of him. But if Java is not the third factor, what is? I suspect that McNealy was close to the mark. It is the technologies underlying Java which he should have cited.
The first respect in which Java differs from the past is that it represents third wave of programming platforms. What makes it so attractive to developers is its write-once, machine independent architecture. And that's down to nothing more complicated than the fact that it's an interpreted language, rather than the compiled languages of the previous generations.
This characteristic can be viewed as a disadvantage as well as an advantage. Being interpreted makes it slower because the operating system has to do more translation before the code can be executed. That's why people are working on just-in-time compilers, new operating systems and Java chips that speed up the translation process on each individual computing platform.
But even without these aids to performance, Java still runs acceptably on today's desktops and servers simply because of the speed of current generation processors. Java couldn't have happened until now because earlier computers weren't powerful enough to produce decent performance without using compiled programs.
Fundamental shift
The other, even more fundamental, trend that Java represents is a shift to componentised software. This development is so recent that it's still in its infancy, and I suspect its immaturity is something that those attempting to build mission-critical distributed applications in Java are even now in the process of discovering.
But the concept of being able to tailor applications to specific user needs from a portfolio of building blocks — and to be able to distribute different parts of the application around various network locations — clearly has the potential to have a revolutionary impact on the nature of computing.
The virtual machine and componentisation are the two crucial technologies that combine with plentiful bandwidth and the Web to create the conditions for the third paradigm of computing that McNealy described. Java is no more than the first popular implementation of those technologies, just as the Apple Mac was the first personal computer driven by a graphical user interface.
And in the same way that Apple lost out to Microsoft, Sun is likely to find itself edged aside by a future innovator that learns from its experience and does a better job of anticipating what the market will buy. It may be two, three, five or more years before it's possible to discern who will really set the agenda in the third wave of computing, but McNealy is set to discover that being first on the starting block is no guarantee of ultimate victory.
more columns