SchemaAware Processing in XSLT 2'0

1
Schema-Aware Processingin XSLT 2.0

• Michael Kay
• Prague 2005

2
Contents

• Motivation why do it?
• Whats different about it?
• Demonstration

3
High-level Rationale

• Stylesheets are written with a particular schema
  in mind
• Schema knowledge ought to be useful
• at design time
• at compile time
• at run time
• Potential benefits
• error detection (and prevention)
• optimization

4
W3C XML Schema

• Highly complex specification
• not universally liked
• many features, but not always the ones you want
• Widely supported by big vendors and users
• More than just validation
• Defines data types
• Supports data binding

5
Schemas and Stylesheets

• Stylesheet can import a schema
• makes the types available for use within the
  stylesheet at compile time
• e.g. matchschema-element(invoice)
• Source documents can be validated
• gives each node a type annotation
• Result documents can be validated
• checks correctness of output
• typed input to next step in pipeline

6
Schema-Awareness is Optional

• Optional for the vendor
• there is an XSLT basic conformance level that
  excludes schema-aware features
• Optional for the user
• you can process unvalidated (untyped) data
• you can mix typed and untyped data

7
Demonstration
Schema
Schema
GEDCOM 5.5
GEDCOM5.5in XML
GEDCOM6.0
XHTML
previousexample
thisexample
8
Conclusions

• Writing schema-aware stylesheets is different
• a different way of thinking
• a different experience of development and
  debugging
• Definite benefits in early error detection and
  diagnosis
• Optimization benefits still speculative