tide floats all boats" attitude, they'
re wasting time and money on infighting. What's up with that?
The increasingly contentious Windows and Java camps are a perfect example. They're in a fierce war. The first casualty: openness. The straw man argument of the year is "Standards bodies take too long. Customers want us to get products to market sooner."
Hello! Has anyone noticed this Internet thing? Produced by standards bodies, wasn't it? Both the Internet community and the communications community have known for a long time how to work with standards bodies, anticipate them, get products to market that embody draft standards and are upgradable, and assure users that the road ahead is not fraught with dead ends. Time for the software industry to place more emphasis on that process and less on shipping beta software as finished products.
The fact of the matter is that neither Windows nor Java is remotely close to being open. That doesn't make them bad, but until Microsoft and Sun turn technologies like ActiveX and
Java over to committees that can really craft the technologies' futures in a consensual manner, I won't call any of them open. Popular, available, inexpensive, and useful -- even extensible -- is not the same as open.
How long can Microsoft say that "Java is a language, Windows is the platform" without becoming irrelevant to a significant portion of its customers who persist in multiplatform computing? Doesn't the company remember when Windows was as immature as Java is now and people still chose it over their installed systems? And the Java crowd is going to have to grow out of its puerile "pure Java" stance to embrace living legacies like Windows.
And what's up with Intel? Here at BYTE we're watching the Slot 1/Socket 7 controversy very closely. The historic, relative flexibility of the Intel architecture has certainly helped Intel become the dominant force it is. Now, Intel seems bent on dictating not just processor architecture but computer architecture as well.
It's a curious tack to
take. A few years ago, we all debated endlessly whether RISC would overtake CISC. That war is over. Intel has brought RISC concepts like pipelining into its architecture. That, and the continued preponderance of integer-based computing, has kept the floating-point kings of the RISC world at bay. If anything, the dual-processor Intel machine running NT has become a very solid alternative to many RISC/Unix workstations. Is Intel that worried about AMD and Cyrix?
We're in the midst of a pendulum swing away from general-purpose computers and toward a greater number of specialized platforms: Web TVs, network computers, PDAs, desktops, uniprocessor servers, quadprocessor servers, and so on. The more Intel owns of the PC architecture, the less able it will be to serve that diversity of needs. If the Sequents and Corollaries of the world had not pioneered symmetric multiprocessing with the 486, would Intel be in a position to turn SMP into a commodity today? No way.
What about emerging technologies like
hand-helds or wearable computers? Are they to be stuck with a one-size-fits-all technology like the Pentium II single-edge cartridge? Or will other chip makers fill their needs? Probably not what Intel had in mind, but it could be the outcome.
Mark Schlack, Editor in Chief,
mschlack@bix.com
Industry Warfare: What's Up With That?
