From RDFaWiki

WARNING: Experimental wiki page. This page outlines concepts that are purely experimental in nature.
WARNING: The work described below is in no way endorsed by the RDFa Task Force, SWDWG or the W3C.
WARNING: The existence of this page does not imply nor express a statement that the concepts here will become a part of RDFa now or in the future.

Contents 1 Introduction 2 Accessing the Design Suite 3 Design Cases 3.1 Mapping prefixes in XHTML1, XHTML2, HTML4 and HTML5 3.1.1 9001 3.1.2 9002 3.1.3 9003 3.1.4 9004 3.1.5 9005 3.2 Extending Reserved Words and Creating New Tokens 3.2.1 9006 3.3 List of Items to Consider

Introduction

The RDFa Design Suite is an area of the wiki and RDFa Test Harness that can be used to explore extensions and new design ideas for RDFa as it evolves. This page should be used for discussion, specification of alternative test cases and should see heavy modification.

Accessing the Design Suite

To access the Design Suite Cases, go to the RDFa Test Harness and select the following:

Test Suite      : "Design Test Suite"
Unit Test Status: "Unreviewed"

Design Cases

All of the design cases below map directly to the RDFa Test Harness design case numbers.

Mapping prefixes in XHTML1, XHTML2, HTML4 and HTML5

This mechanism is needed because the xmlns: prefix mapping mechanism used in XHTML1 is not available in non-XML family languages. The goal is to create an alternate mechanism for mapping prefix CURIEs in RDFa and any other language utilizing CURIEs. There is more detailed information about prefix mappings on the RDFa in HTML4 page.

There are a number of separator mechanisms that have been discussed:

A simple space-separated list of URIs, using an equal sign as the separating token between prefix name and URI:

prefix="prefix1=uri1 prefix2=uri2 prefix3=uri3"

A simple colon-separated list of URIs, using a colon as the separating token between prefix name and URI:

prefix="prefix1:uri1 prefix2:uri2 prefix3:uri3"

A more complex, CSS-like syntax for specifying prefixes using colons as the separating token between prefix name and URI and semicolons as the separating token between each item:

prefix="prefix1:uri1;prefix2:uri2;prefix3:uri3"

The above could also be written as:

prefix="prefix1:uri1; prefix2:uri2; prefix3:uri3"

9001

Establishing @prefix on HTML element

Demonstrates the colon-based prefixing mechanism on the top-level HTML element. This is the mechanism authors would use to specify a prefix mapping for use anywhere in the document.

9002

Establishing @prefix using CSS-like syntax

Demonstrates the CSS-like prefixing mechanism on the top-level HTML element. This is the mechanism authors would use to specify a prefix mapping for use anywhere in the document.

9003

Establishing @prefix using simple '=' syntax

Demonstrates the simple equals sign prefixing mechanism on the top-level HTML element. This is the mechanism authors would use to specify a prefix mapping for use anywhere in the document.

9004

@prefix is allowed on any element

The @prefix attribute should work on any element and follows the same scoping rules as xmlns: (but note that @prefix is /not/ a direct replacement for xmlns. We have discussed specifying prefix in the HEAD of the document in HTML4, such that any prefixes defined in HEAD would be use-able in the rest of the document.

9005

Use of @prefix in HTML4

This test shows the @prefix attribute in use in an HTML4 document.

Extending Reserved Words and Creating New Tokens

In order to support more Microformats-like markup, a mechanism must exist to dynamically extend the list of reserved words or tokens that are recognized by the RDFa parser.

Whichever mechanism is used, it is currently believed that the target document in @profile would contain RDFa statements understood by the RDFa parser to extend the list of reserved words or declare new prefix mappings. For a further explanation of this approach, see the RDFa Profiles document.

9006

@profile attribute can be used to import prefixes and references

It has been mentioned that the @profile attribute could be re-used or extended in the same way that GRDDL uses @profile - to specify a document that defines processing rules for the current document.

List of Items to Consider

• How are XMLLiterals handled in HTML4/HTML5? Should we specify an HTMLLiteral?
  • This is discussed briefly on the RDFa in HTML4 page.
• Should complete URIs be allowed in @rel, @rev, @property, and @datatype so that RDFa can be defined without the use of CURIEs?
• How do we express prefixes in XMLLiterals in HTML4?