W3C home > Mailing lists > Public > public-rdf-in-xhtml-tf@w3.org > October 2008

[DRAFT] SVG Tiny Working Draft Comments from RDF in XHTML Task Force

From: Manu Sporny <msporny@digitalbazaar.com>
Date: Thu, 09 Oct 2008 15:22:08 -0400
Message-ID: <48EE59E0.8060007@digitalbazaar.com>
To: RDFa Developers <public-rdf-in-xhtml-tf@w3.org>

These are draft comments for the SVG WG. Please review quickly and send
any corrections as soon as possible. I will send them off approximately
6 hours from now in order to meet tomorrow's comment submission deadline.


The following are comments from the RDF in XHTML Task Force related to
the latest SVG Tiny Working Draft located here:


The comments revolve around the re-use of RDFa attributes in SVG, which
we are very pleased to see, but feel that there should be specific
direction on how those attributes are used.

Re-use of RDFa attributes should follow RDF in XHTML processing rules

SVG includes several attributes that may be placed on any element, for
the use of attribute-based metadata formats. These include the 'class',
'role', 'rel', 'rev', 'about', 'content', 'datatype', 'property',
'resource', and 'typeof'  attributes. ***SVG makes no specific
requirements about the values for these attributes, other than their
particular value data types, such as a string or a space-separated lists
of strings.*** Other specifications, such as RDFa [RDFA], Microformats
[MF] patterns, or ARIA [ARIA] ontologies,

The current text leaves far too much room for mis-use and abuse of the
RDFa attributes. It would be a shame if authors were allowed to
re-define how a non-RDFa parser may use those attributes in such a way
as to directly conflict, or even worse, create ambiguity with regard to
the current RDF in XHTML parser rules. The RDFa task force went to great
lengths to ensure that the RDFa Syntax Processing[1] rules define clear
behavior when RDFa is used in non-XHTML languages.

Please add text clearly stating that if one re-uses the RDFa attributes
that they follow the same processing rules as outlined in the RDFa
Syntax Processing Rules[1].

@rel/@rev values do not necessarily need to be prefixed

When used with RDFa, the values for the 'rel' and 'rev' attributes must
be a CURIE [RDFA] (i.e., a prefixed string, such as 'cc:license' to
indicate a Creative Commons license), while the values may simply be
from a set of specific keywords for Microformats. These formats may be
used independently, or in combination if the keywords do not clash.

The @rel/@rev values in RDFa can either be a reserved word or a CURIE as
defined in the RDFa Syntax document[2]. Did the SVG WG mean to allow
specific keywords (reserved words) for Microformats and not for RDFa? If
so, why the distinction? Why not follow the current RDFa spec, including
reserved words, more closely? It would make the job of RDFa parser
authors much easier as there are less special cases to consider when
creating their parser. The current language requires all current RDFa
parsers to strip out all reserved word processing in order to conform to
SVG, which would lead to two increasingly divergent types of RDFa parsers:

1. Parsers that parse SVG RDFa.
2. Parsers that parse "XHTML+RDFa 1.0".

In addition, the document seems to insist that terms must be of the form
"x:y", instead of also allowing things like "license" (note that there
isn't a preceding colon before a reserved word). It would be advisable
not to limit @rel/@rev values to colon-only forms. We are currently
working on a method to specify reserved words that will not need
preceding colons.

Please do one of the following:

* Adopt the current RDFa processing rules as-is.
* Define a set of reserved words that should be used in SVG Tiny.
* Do not rule out the ability to use non-colon-prefixed reserved words.

@role should follow rules defined in XHTML 1.1 Role Module

This is not an official comment from the RDF in XHTML Task Force and
will probably be mentioned by the XHTML WG. Mark, Shane and Steven were
on todays RDFa telecon and had issues with the lack of specifics as to
how an author could use @role. Just a heads up that they would like to
see it clearly stated that use of @role should follow the specifics
outlined in the XHTML Role Attribute Module[3]. We want to make sure
that authors are not under the false assumption that they can put
whatever they want to in the @role attribute.

-- manu

[1] http://www.w3.org/TR/rdfa-syntax/#sec_5.5.
[2] http://www.w3.org/TR/rdfa-syntax/#relValues
[3] http://www.w3.org/TR/xhtml-role/

Manu Sporny
President/CEO - Digital Bazaar, Inc.
blog: Bitmunk 3.0 Website Launches
Received on Thursday, 9 October 2008 19:22:50 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:50:29 UTC