Prefixes in RDFa for HTML

1. Introduction: Background and state of play

RDFa was originally defined in a single document which assumed/defined an XHTML-embedding for it RDFa in XHTML: Syntax and Processing, W3C REC 2008-10-18. The next generation of RDFa, which separates a matrix-independent core from matrix-specific embedding specifications, is currently under development, with WDs for RDFa Core 1.1, W3C WD 2010-10-26, XHTML+RDFa 1.1, W3C WD 2010-11-09 and HTML+RDFa 1.1, W3C WD 201-01-13.

RDFa uses CURIEs extensively as the values of attributes. The non-empty prefixes in those CURIEs are interpreted relative to

• XML namespace prefix bindings, established by xmlns:... pseudo-attribute: [RDFa in XHTML (2008)] (and [RDFa 1.1], optionally)
• RDFa prefix bindings, established by prefix attribute: [RDFa 1.1] only
• RDFa prefix bindings, established indirectly from profile documents referenced by profile attributes: [RDFa 1.1] only
• RDFa prefix bindings, established implicitly ("intrinsically known to the parser"): [RDFa 1.1] only

Preserving the prefix bindings arising from xmlns:... in non-XML-mode XHTML5 processing adds complexity to implementations and, it is argued by Ian Hickson, the use of prefixes confuses some users.

The result is an issue raised by Hickson which has garnered two change proposals:

• One from Hickson proposing to drop support for prefix binding altogether;
• One from Toby Inkster which proposes the status quo, with additional documentation

2. Comment

Hickson's change proposal focusses on arguments against any use of a prefix-declaration-and-expansion mechanism, including fragility and existing implementation problems. There is little discussion of the requirements for the use of CURIEs in the first place, except indirectly: there is an assertion that using full URIs everywhere is not problematic, on the basis of a citation of a study, which turns out to have had seven subjects looking at microData examples, with very little (no?) authoring or indeed being asked about URIs at all, in which Hickson comments:

One thing we weren't trying to test but which I was happy to see is that people really don't have any problems dealing with URLs as property names. In fact, they didn't even complain about URLs being long, which reassured me that microdata's lack of URL shortening mechanisms is probably not an issue.

Inkster's proposal focusses on examples of existing uses of CURIEs in RDFa in the wild. He does include some suggestions for change in the area of increasing the prominence of examples of full URI use in the RDFa specs. Inkster also points out the divergence which would arise between RDFa for XHTML and RDFa for HTML if Hickson's change is approved. He doesn't discuss the CURIE requirement issue either.

Neither proposal contains any concrete evidence about the utility, or lack of it, of prefixes for authors, or the importance, or lack of it, of that utility to authors and consequently to uptake.

This issue also obviously interacts with the HTML WG's long-standing Distributed Extensibility issue: if xmlns:... prefix bindings have to be processed for use in CURIEs, then it becomes harder to argue against allowing their use in element and attribute names. . .