Re: RDDL proposal from Sandro Hawke (LinkTypes/NMTOKENS issue)

> Sandro Hawke wrote:
> >    <a rel="http://www.rddl.org/purposes#validation"
> 
> Sadly, this is invalid HTML in some versions. rel is defined as 
> containing CDATA in HTML4, but I was told by Murray Altheim that this 
> is a mistake. In later (and earlier?) versions, it was "fixed" by 
> changing it to NMTOKENS. :-(

Oh, Nasty!  I see now your old debate about this [1].  From HTML 3.2
through XHTML 1.0, LinkTypes was CDATA [2], then in "modularizing"
XHTML 1.0, it got changed to NMTOKENS [5].  The justification for this
appears to be that it was SGML names in earlier HTMLs [3].  (IMHO, the
relaxing to CDATA with 3.2 was a logical progression, not a bug.)  The
text even in the current XHTML 2.0 draft (as you said) still suggests
a list of URIs would be fine. [4]

Ironically, this proposal was my first version to actually go beyond
NMTOKENS.  My earlier ones used an explicit namespace-like import
step, to get the relationship names into NMTOKEN space.  But that
seemed unnecessarily complicated for the RDDL challenge, so I dropped
it.  I guess for people who (for some reason!) want to use HTML 2.0 or
XHTML 1.1, I'll reintroduce my import mechanism.  Hopefully XHTML 2.0
will do the right thing here and properly use RDF for identifying
relationships.

   -- sandro

[1] http://lists.w3.org/Archives/Public/www-archive/2001Apr/0010
[2] http://www.w3.org/TR/html4/sgml/dtd.html#LinkTypes
    http://www.w3.org/TR/html401/sgml/dtd.html#LinkTypes
    http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd
    http://www.w3.org/TR/REC-html32#dtd
[3] http://www.w3.org/MarkUp/draft-ietf-html-relrev-00.txt
[4] http://www.w3.org/TR/2002/WD-xhtml2-20021218/abstraction.html#dt_LinkTypes
[5] http://www.w3.org/TR/xhtml-modularization/dtd_module_defs.html#dtdentry_LinkTypes.datatype

Received on Thursday, 26 December 2002 13:53:43 UTC