(March 20, 2002) - The co-inventor of XML, Tim Bray, has been stirring up a lively debate among Internet technologists. The reason? He thought it might be fun to make a public presentation in which XML - and indeed every other Internet technology of the past and present - was "ranked" according to the simple criterion: does it actually matter?

Widely recognized as an expert in information search and retrieval from large databases, and former cofounder of the Open Text Corporation where he introduced one of the first commercial search engines, Bray tends to know whereof he speaks. So his verdict - which was that SQL/RDBMS, Unix/C, Open Source, PC Client, WWW, Java, and XML have all been proven to matter...while OODBMS, 4GL, AI, VRML, Interactive TV, Ada, and SGML have been proven not to - is noteworthy.

Arguments about which are the "matter" technologies and which are the "don't matter" ones are always likely to break out at intervals. So it's no surprise that it should do so now in the XML arena. Some experts would say that the key differentiator for XML is that it has always had the standardization process full-square behind it, elevating it on a level with, say, the US television standard and the GSM mobile phone standard. Others point out that the dominant factor seems to be that people will adopt a technology if they believe that everyone else is going to do so, but won't if they don't.

Bray's rating of SGML as a "doesn't matter" technology is unlikely to endear him to some. Particularly as XML 1.0 derives directly from SGML, of which it is a subset! And many advocates of SGML would insist that it is more sophisticated than its younger cousin.

XML-J Industry Newsletter turned to Joshua Allen - whose calm analysis of the i-technology scene can usually be depended upon 100% for its acuity and accuracy alike. Allen says: "I have to side with Tim on this one."

"When Tim says 'matter,' he is talking about the kind of broad impact that the PC has had," Joshua Allen continues. "Steve Ballmer has said that XML is 'bigger than the PC,' so if the PC matters, then obviously XML matters."

What about SGML, is Bray snubbing it when he says it doesn't "matter" in the same way? Not so, thinks Allen.

"This is not necessarily a snub of SGML," he says. "The XML team took many concepts from SGML and made them accessible to a much broader audience. Just as the PC took computing power out of the data center and put it in the hands of the masses, XML has taken SGML power and put it in the hands of the masses."

Allen is characteristically even-handed when he concludes: "Big data centers are still alive and useful, and so is SGML - and the PC and XML owe a debt to both. But the XML and the PC occupy a 'sweet spot' of influence that combines low barriers to entry, wide interoperability, and a HUGE install base."

XML-J IN thanks Allen for his excellent two cents' on this. Tim Bray must be hoping that Charles F. Goldfarb will be half as generous and understanding of his point of view when he catches up with this story!

• One of XML's Founding Fathers Joins Sun