Four years after the publication of XML 1.0 as a W3C Recommendation, Edd Dumbill opens the XML Europe 2002 conference by questioning whether XML has become the "SGML on the Web" which it was meant to be.

XML has become a widespread technology:

"XML has achieved more than we could believe and is even ridiculously widespread to the point that we find it in the configuration files of our PC. It's been used to integrate to web web cell phones and set top boxes and is even been taught in the universities. "

But the stage of marketing hype has passed and the hype has now moved on to Web Services:

"One the nice things about this year is that we've pushed through the hype bubble. In 1999 and 2000 people had been expecting XML to produce some magic economical effect and that's been counter-productive to a sensibly measured way of developing things. XML has now proved that it was an evolutionary thing."

One of the foundations of XML is agreement, although this is not always visible in discussions between experts:

"XML is based on the agreement on a core syntax and on standard tools and this is infectious: many groups share their developments and DTDs describing things as diverse as recipes get propagated, however the heat of the discussions shows also a good deal of disagreement. But disagreement itself may be the proof that there is room for more than one approach and this is healthy since one could have thought that the agreement on XML could have cut everything down."

But disagreement is sometimes needed to insure that different sets of requirements are met:

"We need to insure that all users of XML are carried forward as we are getting on. The fact that there are different standards bodies is a good thing since they are driven by different requirements. The W3C is driven in a direction which is OK for some applications but is not serving everybody, and there are people considering XML as documents and not as data. Happily, there is a community and now standard groups that insure that there is a standardized future for these applications."

And the Semantic Web appears to be the next step toward really putting SGML on the Web:

"One of the current buzzwords is the Semantic Web and I am very excited by this vision: the Semantic Web is about making the web machine processable and in a way the Semantic Web is still trying to put SGML on the web. It's just that putting SGML on the web was not quite the same thing that we thought it was when we started out."