Hard choices ahead for developers

The subject of discord was software and my opinion of the quality of his company's programming talents. I had pointed out in a column that, despite all its self-serving public relations, Microsoft had utterly failed to distinguish itself as a developer of really great software.

I went further, contending that the truth of the matter is that Microsoft had a lousy track record of turning out big hits with first versions. Anybody remember the 1.0 incarnations of Windows? How about NT? Or Microsoft Bob?

But I also wrote that Microsoft does an absolutely terrific job reinterpreting the design of its software in response to customer feedback. That's no inconsequential talent--and a central reason for the improving usability of its products.

Time and again, Microsoft has revived prospects of product lines that got off to inauspicious starts. (Of course, no amount of user feedback could rescue Bob, an interface designed for cretins. The product was so awful that Microsoft was forced to pull the plug--even though it was the pet project of one Melinda French, soon to become better known as Melinda Gates.)

Maritz wasn't in the mood. Regardless of cockeyed observations by certain computer columnists, he said, Microsoft prided itself on making "great software," and the company's record stood for itself.

Had time stood still, this debate wouldn't be particularly significant. But it's not 1996 anymore. The company is shifting toward a software-as-a-service model, and that means many developers will also need to shift their efforts from creating products for the desktop to finding better ways to bring together XML-linked data over networks.

Microsoft CEO Steve Ballmer has been busily wooing developers all across the United States and beyond. He knows the success of his company's Web services strategy depends on getting developer support. But the fount of imagination and creation in the computer industry doesn't work with top-to-bottom diktats. Independent software developers--not Microsoft--are the ones who have usually come up with the great stuff.

That's why this is an especially tricky time in the history of small developers. Should they trust Microsoft or stay true to their instincts?

I suppose some developers will decide to sit this one out or seek other allies, but it's going to be hard to pretend that Microsoft doesn't exist in this evolving world order. A few years ago, developers had a stark choice: Support Microsoft's Component Object Model or the Common Object Request Broker Architecture favored by Java supporters.

Those were proprietary protocols and specs. This time around, Web services are going to be "open" standards. But things will run more smoothly if you use the special sauce cooked up in Redmond. And so, Microsoft is getting Visual Studio.Net and .Net Framework tools into the hands of developers as it starts pushing a software-development architecture it calls the Global XML Web Services Architecture.

Microsoft will eventually submit this to a standards body to get universal benediction. In doing so, the company will be able to argue that the specs are built on XML and SOAP (Simple Object Access Protocol), two of the emerging standards behind the shift to Web services.

OK, but is this necessarily the best thing for small developers? I'll weasel out on this one with the old columnist crutch: Only time will tell. It's not at all clear that SOAP is necessarily better than XML-RPC. The support of the larger companies in the industry, like Microsoft, may very well decide that issue in SOAP's favor. But that doesn't mean it's the better choice .