Coming of Age in Cyberspace: Births, Deaths and Milestones at SGML/XML '97: Conclusion

January 10, 1998

Seybold Report on Internet Publishing
Vol 2, No 5
January, 1998

The closing keynote by C. Michael Sperberg-McQueen, co-editor not only of XML but of the Text Encoding Initiative and now co-chair of the GCA conference, has become a much-anticipated annual event that summarizes where the industry has been and where it is likely to go in the following year. His closing question to the conference, "Now that we will have XML apps on every desktop and every webtop, what will we do with it?" is a good one to ponder for the coming year.

Steve Sklepowich, product manager for publishing technologies at Microsoft, sees low-hanging fruit in the current SGML docbases—all those gigabytes of technical documentation that will now have easy and effective access to the Web.

We have some answers of our own in terms of applications and the refinement of the standard itself.

  • Clearly, nontext applications for XML, such as Open Financial Exchange and XML/EDI, will proliferate.
  • We will see more development of middleware and the APIs required to reduce time spent by XML programmers.
  • We will see the creation and adoption of a non-proprietary, application-independent style language for XML that is richer and more flexible than CSS, although we don’t expect to see it adopted by the market until 1999. CSS will prepare a style-sheet mentality that will pave the way for XSL, much as HTML created a markup-aware user population that is now receptive to XML.
  • Concurrently, we will see development of a "new WYSIWYG" style of editing with interactive, direct representation of formatting applied to structured text through a style sheet editor or directly onscreen.
  • There remains, as Tim Bray, co-editor of the XML specification, put it, "a hole in the authoring application arena." The current generation of authoring tools is just too difficult for the average user to master. We see some continuing innovative work by yet more small European players, which encourages us that the hole can be plugged, but the announcement after the show that Office ’98 applications will be able to save into XML does little more than whet our curiosity and our appetite for creative approaches to "fun with structured text."
  • We will get some form of XML-data that will give us new methods of expressing architectural forms, inheritance and strong-data typing. XML-data will greatly accelerate the integration of structured, application-independent markup with mainstream computer science. This process will be shadowed by tension between those who would trash the SGML-specific schema designs of the past in favor of a brave new beginning and those with cooler heads. Prudence dictates preservation of a strong link with established practice, both for the experience and guidance embodied in a decade of DTD design and for the political protection of the ISO-validated standards process. But there are those who would just as soon ditch DTDs for validation and replace them with something more akin to database schemas.
  • RDF-related efforts will pioneer new ways to represent knowledge through directed graphs in structured markup that will do much more for our understanding of text than just locating a few cool Web pages. The application of entity-relationship theory to structured text will open up new areas of knowledge representation and experimentation that will be relevant beyond publishing and beyond transactional messaging. It will take data representation out of the shadows and into the limelight with consequences that are still too poorly understood to predict.