Facet Segregation, AspectXML, Web Publishing, and XML 2006

The XML 2006 conference proceeding were due today

"The Essence of Declarative, XML-based Web Applications: XForms and XSLT"

The term Facet Segregation is used in this paper to identify a methodology for separating and abstracting individual aspects of XHTML and its derivative, related dialects into single-purpose vocabularies and using XSLT to compile more expressive languages from them. It can be considered an application of AspectXML for the composition of web dialects. It extends the well-established method(s) of separating content from presentation (used by DocBook XSLT stylesheets, for instance) with both Gleaning Resource Description Dialects of Languages (GRDDL) and the value proposition presented in this paper. It is within this larger picture that the value proposition of separating presentation markup from more specific application logic is proposed.

In the paper I demonstrate how XUL and a simple vocabulary for high-level user interface components (at a level more abstract than the widgets and controls commonly associated with user interface languages) can be used to compose a user interface for editing Atom entries. From this abstracted user interface document, an XForm is generated using an XSLT stylesheet that matches XUL, XHTML, XForms and the elements in the abstract vocabulary. I chose not to drill down too deeply into the example and spent more time describing how the value of using XSLT to generate XForms in this way fits in the larger picture (i.e, web architecture, rich web application backplane, semantic web architecture, and document composition).

I also attempt to describe a general framework of using XSLT to facilitate the extraction of various representations of the same document each of which facilitate a particular facet of web content:

  • Presentation
  • UI Behavior
  • Semantics or 'Faithful Rendition'

Faithful rendition is a term used in the GRDDL specification that has grown on me and represents the authors intent of the 'meaning' of a document as determined by the application of a GRDDL transform:

By specifying a GRDDL transformation, the author of a document states that the transformation will provide a faithful rendition of the source document, or some portion of the source document, that preserves its meaning in RDF.

The general idea (from 500,000 feet) here is that the means is the same but to different ends and enables an unprecedented level of richness for web content - when used in combination. I use the term Facet Segregation to describe this general approach. I ended up prefering this to 'Modality Segregation'. There is some correlation with Aspect Programming and AspectXML. The diagram below is my attempt to capture this:

The (Docbook) in the diagram identifies the most common usecase for the oldest of the three 'mechanisms' (seperation of content from presentation - typically facilitated by Docbook XSLT stylesheets).

[Uche Ogbuji]

via Copia