RDFa in XHTML: Syntax and Processing Specification Errata

Last modified:

$Date: 2009/11/19 17:43:38 $

This document records known errors in the document:

http://www.w3.org/TR/2008/REC-rdfa-syntax-20081014

The latest version of the RDFa in XHTML: Syntax and Processing specification is at:

http://www.w3.org/TR/rdfa-syntax

Copyright ©2001-2008 W3C^® (MIT, ERCIM, Keio), All Rights Reserved. W3C liability, trademark and document use rules apply.

---

Comments on this specification may be sent to www-html-editor@w3.org.

Known errors

1. Section 5.5 step 6 - The text of step 6 reads "If in any of the previous steps a [new subject] was set to a non-null value, it is now used to provide a subject for type values;". This is somewhat misleading. By the time step 6 is reached, there will ALWAYS be a [new subject]. What this step is really about is that if there is an @typeof attribute, it is used to define the types for the [new subject]. This is clearly described in the explanatory text.
2. Section 4.1. Document Conformance - In the future it is possible that RDFa will also be defined in the context of HTML. Consequently document authors SHOULD use lower-case prefix names in order to be compatible with current and potential future processors.
3. Section 5.5 step 9 - The text of step 9 reads in part "The value of the [XML literal] is a string created by serializing to text, all nodes that are descendants of the [current element], i.e., not including the element itself, and giving it a datatype of rdf:XMLLiteral.". For the avoidance of doubt, this means in part that the current default namespace of each descendant element MUST also be included in the emitted XML literal.
4. Sections 5.4.5 and 7 - The text indicates that the prefix '_' is reserved and is used to create / reference blank nodes (bnodes). Because this prefix is reserved, authors SHOULD NOT declare a mapping for the prefix '_' and conforming processors MUST NOT incorporate such a prefix mapping into the 'list of URI mappings' as defined in section 5.5.
5. Section 5.5, step 9 reads in part "The value of the [XML literal] is a string created by serializing to text, all nodes that are descendants of the [current element], i.e., not including the element itself, and giving it a datatype of |rdf:XMLLiteral|." Unfortunately, the requirements for such a serialized string are not completely clear when it comes to XML Namespaces. For the avoidance of doubt, in future versions of the RDFa Syntax Recommendation, we expect to indicate that the value of the XML literal SHOULD be serialized Exclusive Canonical XML as defined in [XMLEXC14N]. We advise current implementations to start supporting this use model now, since other participants in their XML tool chain may expect consistent, canonical XML.

Clarifications

1. Rules for evaluating CURIEs and SafeCURIEs are defined throughout the Recommendation. For the avoidance of doubt, these rules require at least the following:
   • The set of prefix mappings to use when expanding CURIEs to IRIs is provided by the current in-scope prefix declarations of the element on which the CURIEs are found (see section 7 and section 5.5 step 2).
   • The URI to use when the 'default' prefix is referenced from a CURIE (e.g., ':next') is '|http://www.w3.org/1999/xhtml/vocab#' (see section 7). An RDFa Processor MUST expand CURIEs prefixed with only a colon using this prefix mapping.
   • There is no mapping defined when there is no prefix AND no colon specified (see section 7). Therefore, an RDFa processor MUST NOT expand non-prefixed CURIEs (however, see reserved terms below).
   • The prefix '_' is reserved and is used to refer to document-local blank nodes (bnodes) - see sections 5.4.5 and 7. RDFa processors MUST ignore any declaration of a mapping for that prefix.
   • The reserved terms (defined in section 9.3) are ONLY processed on the attributes rel and rev. All other attributes that take CURIEs or SafeCURIEs as values MUST NOT treat reserved words specially. The datatype definitions are in section 9.1, and the attribute to datatype mappings are in section 9.2.
   • Prefix declarations are subject to the syntactic restrictions defined in the Namespaces in XML Recommendation (see section 5.5 step 2). These restrictions include prohibiting the declaration of a prefix mapping for the special prefix 'xmlns' and limiting the declaration of the special prefix 'xml' to map to its pre-defined namespace URI (see the Namespaces in XML Recommendation, section 3). RDFa Processors MUST behave as if these restrictions are enforced.
   • Section 9.3 defines a collection of "reserved word" values. These values are only interpreted in the context of @rel and @rev. @rel and @rev have been historically defined to be case-insensitive. Section 9.3 does not indicate whether the reserved words are case-sensitive or case-insensitive. However, Section 5.4.4 does mandate that the use of these values are required to be mapped to URIs in the XHTML Vocabulary via fragment identifiers, and fragment identifiers are by definition case-sensitive. For the avoidance of doubt, a Conforming RDFa Processor MUST treat the reserved word values as case-insensitive on input, and MUST transform the values to lower-case when mapping them to URIs in the XHTML Vocabulary.

References

[XMLEXC14N]

"Exclusive XML Canonicalization - Version 1.0", W3C Recommendation, J. Boyer et al., eds., 18 July 2002.
Available at: http://www.w3.org/TR/2002/REC-xml-exc-c14n-20020718/