Date: Mar. 9, 2001
Link: http://xml.oreilly.com/news/xmldevcon_0201.html
Source Author or Organization: Frank Willison, Editor In Chief, O'Reilly & Associates

In this article, the author returns from XML DevCon Europe to report on the "two major directions of XML": abstraction and metadata.

Progress in computing in general has been characterized by a continual move towards ever greater abstraction; in hardware we've journeyed from mainframes to embedded and wearable devices. A similar evolution has occurred in the data arena, with XML emerging as "the ASCII of the twenty-first century".

Rather than dealing with text strings and angle brackets, the modern user will see data transformed and organized into infosets manipulated by XML utilities. These "infosets" will be "piped" to other utilities that can then act upon the data (somewhat analagous to the way Unix programs can act upon and transfer data via Unix pipes).

The author gives a summary of the expected process: the XML parser will receive a character stream and build an infoset, XML Schema can act on the resulting tree structure to assign types, one infoset can be mapped to another with Extensible Style Language Transformations (XSLT), the schema-validated infoset can be utilized by Xinclude and XLink, and the final infoset can be queried by XML Query and distributed by XML Protocol, a new specification that builds upon Simple Object Access Protocol (SOAP).

The success of this abstraction will require rich metadata. The author describes every Document Type Definition (DTD) and XML Schema as "a coherent set of abstractions for a particular purpose". He found the proliferation of methods to describe data surprising, and lists as particularly worthy of further research: Resource Directory Description Language (RDDL); Universal Description, Discovery, Integration (UDDI); and Web Services Description Language (WSDL).