W3C wants the Web to be a big database

NEWS The World Wide Web Consortium (W3C) published updates to a group of interlinking specifications that recommend uniform ways to retrieve information from XML (Extensible Markup Language) documents. The publications include two "last-call" drafts and a brand new one. The updates were published last week in the run-up to the XML Europe 2003 Conference & Exposition, which opened in London on Monday. The show brings together representatives from high-tech companies, standards bodies and users' groups interested in XML. XML forms the basis of technologies designed to help computer systems communicate more coherently with one another to deliver Web services, an emerging enterprise computing trend. The W3C recommends XML for structuring data, and the task of making XML behave more like a relational database falls to the organisation's XML Query working group. "How do you make traditional database languages like SQL (Structured Query Language) work with XML?" asked W3C spokeswoman Janet Daly in an interview with CNET News.com. "The XML Query working group has been putting together a framework of documents that provide the technical answer to that question, so that XML documents can start to look like parts of one massive database." Members of the W3C's XML Query working group include Microsoft, Oracle, IBM and DataDirect. The ten drafts address various related W3C projects, including XML Query (XQuery), which establishes how to search XML documents; XML Path Language (XPath), which shows how to label discrete parts of an XML document; and Extensible Stylesheet Language Transformations (XSLT), which allows for the translation of one kind of XML document into another, or into a non-XML document.

• XQuery 1.0 and XPath 2.0 Data Model defines what kinds of information applications of XSLT, XPath and XQuery can use.

Generally speaking, a data model would dictate that the addition of integers, e.g. "1+2," would yield only another integer, i.e. "3." The new Data Model draft is based in part on the W3C's Information Set specification, recommended in October 2001. It also makes reference to the W3C's XML Schema recommendation, which supplants the Document Type Definition (DTD) in telling the computer what XML-based language it is reading and how to interpret it. The W3C advanced the data model to "last call" status, meaning that it considers its basic work to be fundamentally complete and ready for limited implementation and extensive public comment. Barring unforeseen problems, the specification should advance to "candidate recommendation" status by the end of June. If the data model becomes a candidate recommendation, the other working drafts updated on Monday would advance to last call, Daly said.

• XQuery 1.0 and XPath 2.0 Functions and Operators, the other draft to reach last call on Monday, expands ways of defining data types within XML. XML Schema established simple data types, such as a number, a temperature, a time and a date. But Functions and Operators allows for the combination of those simple Schema data types and the ordering of them in a sequence.
• XQuery 1.0 and XPath 2.0 Formal Semantics is a draft that establishes mathematical and quasimathematical ways of assigning precise meaning to data model terms shared by XPath and XQuery.
• XML Path Language (XPath) 2.0 expands on Version 1.0, recommended in November 1999, by making the draft a subset of XQuery 1.0.
• XSL Transformations (XSLT) 2.0 changes from its Version 1.0, also released in November 1999, in working better with XPath 2.0 and with Functions and Operators, and by sharing a data model with XPath 2.0.
• XQuery 1.0: An XML Query Language uses the structure of XML to make queries across different kinds of data, whether or not they are originally written in XML.
• XML Query Use Cases spells out day-to-day problems that XQuery drafts could be expected to solve, while XML Query (XQuery) Requirements specifies the working group's goals for the technology.
• XSLT 2.0 and XQuery 1.0 Serialisation is the only brand-new working draft among those published on Monday, though it has roots as part of the XSLT recommendation. Serialisation expresses a value in a string or element that computers can read, store and reuse.
• XQuery and XPath Full-Text Requirements is a draft that sets down requirements that the working draft should meet. In this case, the draft will detail how to use text searches on XML documents.

Details of the drafts can be found at the W3C's Web site.

---