XTAN — XML Transform-as-Needed

What is it?

XTAN allows XML software to be written expecting input in one language and then successfully handle inputs written in other (similar) languages. The other languages can be "competitors", other languages written by other people for a similar purpose, or they can be later versions or variant versions of the original language.

XTAN may be the silver bullet to the forward compatibility problem. In decentralized systems like the Web, where we can't upgrade all the software at once, deploying any new version of a language requires some level of forward compatibility. In many case, the available level of forward compatibility (such as provided by HTML's ignore-unknowns rule) is really not enough. XTAN gives us a lot more.

XTAN is a fairly simple idea. You get an XML document. If it doesn't match the schema you have in mind, use the XML namespaces as web addresses to retrieve annotated schemas. Annotations in those schemas provide transform information which allows you to rewrite the document to match the language you're looking for. While you're re-writing, notice that some transforms are "lossy". Keep track of the losses, and see how bad they are when you're done.

For more details, follow the links below


XTAN was proposed to the Rule Interchange Format (RIF) Working Group in February, 2008, and was well received as a potential solution to addressing its extensibility requirement. In May, however, RIF decided on an XML syntax where all dialects can be mechanically translated to RDF. This may allow RIF to use an RDF-based approach to extensibility, instead of XTAN. That approach needs to be explored in more detail, then RIF will have to decide which approach to take.

In April, 2009, RIF-WG decided not to standardize in this space. It was, in a sense, out of scope for the WG. It wanted something like XTAN, but was not quite the right group to develop such a technology. If you're interested in taking up or supporting this kind of work, please contact Sandro.

See Also

Feel free to send e-mail to sandro@w3.org. Let him know if you want to be on a mailing list. (If enough people do, he'll make one.)

Open Issues

Namespace documents. How do we get to the XML schema from the namespace URI?

Sandro Hawke, W3C
$Id: Overview.html,v 1.12 2009/04/21 18:53:10 sandro Exp $