XML News from Monday, October 30, 2006

Uche Ogbuji has released 4Suite XML 1.0, an open source "comprehensive library for XML processing. It is implemented in Python and C and supports XML (SAX-like and DOM-like), XPath, XSLT, RELAX NG, XUpdate, XInclude, XPointer, and more."

The W3C Device Independence Working Group has posted a second last call working draft of Content Selection for Device Independence (DISelect) 1.0. They've also split out the XPath parts into a separate document, Delivery Context: XPath Access Functions. Finally they've started work on a primer about all this, though it's mostly empty at this point.

According to the abstract, "This document specifies a syntax and processing model for general purpose content selection or filtering. Selection involves conditional processing of various parts of an XML information set according to the results of the evaluation of expressions. Using this mechanism some parts of the information set can be selected for further processing and others can be suppressed. The specification of the parts of the infoset affected and the expressions that govern processing is by means of XML-friendly syntax. This includes elements, attributes and XPath expressions."

That sounds unobjectionable, but what the working group is really proposing is XML markup that can be added to a page to indicate which devices certain content is appropriate for. For example, this sel:if element says that the image should only be displayed if the user's device supports color or has a window size wider than 500 pixels.

<div sel:expr="dc:cssmq-width('px') &gt; 500" 
    & dc:cssmq-color() > 0" >
  <object src="picture.png"/>

This feels more than a little like presentation based markup. This is very much like using JavaScript or server side programs to identify different browsers and send them content tailored specifically to them. This syntax is definitely easier-to-use, and more powerful than the various JavaScript and server-side hacks people use today; but should we be doing this at all? Whatever happened to the vision of sending browsers XML documents with appropriate stylesheets and letting the client decide how to best present it? The thing that bothers me the most about this proposal is that the syntax mixes the presentation information straight into the document, rather than linking to it from a separate hints sheet. In many ways, this document seems to reflect a belief that the W3C has been going down the wrong road for the last eight years in attempting to separate content from presentation.

Bill de hÓra and Joe Gregorio have posted the eleventh public working draft of The Atom Publishing Protocol, a REST-based system for communicating with weblog servers. Significant examples here seem to be related to the handling of collections, media collections, and the their metadata. Major changes in this draft include:

Alex Milowski has released Atomic, a Firefox extension that supports the Atom Publishing Protocol (APP). APP is a RESTful protocol for publishing web logs and similar content repositories.

Brendan Taylor has released atom-tools, a Ruby library that "provides an easy way to manipulate entries and feeds along with the HTTP bits needed for an APP client or server. It includes a small APP suitable for testing clients.

William F. Hammond has posted gellmu 0.8.3, "a LaTeX-like way to produce article-level for online display in the modern, fully accessible, form of HTML extended by the World Wide Web Consortium's Mathematical Markup Language (MathML)."