xmlhack: When validation becomes fuzzy

Rick Jelliffe outlined at XML Europe 2002 that there are many stages in the lifecycle of an XML document "When Well-Formed is too much and Validity is too little" and presented a validation process that can be seen as a suite of small validation steps, ranging from documents that are not well-formed to documents that are more than valid.

Making a comparison between the validation processes of SGML and XML, Jelliffe demonstrated that the notion of validation is relative and evolving, and that "well-formed" and "valid" are just two steps in a validation process that can be split into an array of small validation steps.

On one side of the range, documents which are "less than well formed" are common during the process of editing XML documents and, alas, also on the web. However, most of them already have enough structure to be "weakly validated" and having infeasible structures detected at this stage. On the other side, gradual validation may be applied to a document depending on its stage in its lifecycle, using for instance Schematron's phases.

This notion of fuzzy or progressive validation should help users editing XML documents, and is implemented in the Topologi Markup Editor, an innovative XML editor which should be released this quarter.

Other stories:

James Clark has now added a switch to Jing to allow "feasible validity" checking against RELAX NG schemas. This is useful for validating incomplete documents, where you want to check that you have not made any mistakes so far.

For example, you may be marking up some text and intend to add the "required" metadata later: you just want to find out if there are problems (elements names wrong, etc) in the elements you have done. You don't want to be bored by "errors" which are really due to your workplan, nor (worse) have the validation fail and stop before it gets to the elements you have been working on.