Getting down from the tree
08:53, 28 May 2002 UTC | Eric van der Vlist

XML has developed a tree-centric vision of the documents which is insufficient for some applications. Patrick Durusau urged us at XML Europe 2002 to "get down from our trees" and explore "Concurrent Markup for XML Documents", a XML based technique to define several intermixed tree structures for the same XML document.

Documents often have more than one structure. Durusau presented the example of poetry, which has a both a "technical" structure in term of verses or lines and a text structure in term of sections which often overlap. Explaining that extensive work had been done by Michael Sperberg-McQueen and Claus Huitfeldt on the subject, Durusau notes that this work goes beyond what is possible in a well formed XML document and that "it is tenatively suggested that further research on the question of whether structured markup requires a single tree or if viable alternatives can be devised within the SGML/XML framework is merited".

The solution presented by Durusau is to take the whole markup approach upside down and to consider the markup of the different views as metadata associated with the text split into words. This markup is thus expressed as XPath expressions defining for each word the "place" where it fits in the different hierarchies. Although very verbose (Durusau mentioned an expansion factor of ten) this approach has the benefit of using well-formed XML documents and common tools such as XSLT and XPath.

Related stories:

xmlhack: developer news from the XML community

Front page | Search | Find XML jobs

Related categories