W3C home > Mailing lists > Public > public-swbp-wg@w3.org > October 2004

Re: XHTML 2.0 Metadata proposals

From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
Date: Mon, 04 Oct 2004 12:10:47 +0100
Message-ID: <41612FB7.5060407@hplb.hpl.hp.com>
To: Phil Tetlow <philip.tetlow@uk.ibm.com>
CC: SWBPD list <public-swbp-wg@w3.org>

Phil Tetlow wrote:
> 
> 
> 
> Maybe I’m missing a trick here, but I’ve just read the XHTML 2.0 proposal
> metadata sections and have come away feeling somewhat disappointed. If I
> have got this right, the idea is to include a fixed set of new tag
> properties to allow specific types of in-line contextual ‘metadata’
> augmentation?

I think you have missed a trick ...

http://www.w3.org/TR/2004/WD-xhtml2-20040722/mod-meta.html#s_metamodule

e.g.

<head>
     <meta property="author">Mark Birbeck</meta>
     <meta property="created" content="2004-03-20" />
</head>

the tag set is fixed, e.g. meta and link being the two key elements, 
property being one of the key attributes. But the metadata properties 
are open, the value of the property attribute is a qname which can be 
interpreted like an RDF property name (i.e. using the RDF QName to 
URIref mapping, concatenation).

as an example

   <head xmlns:dc="http://purl.org/dc/elements/1.1/">
     <meta property="dc:creator">Mark Birbeck</meta>
   </head>

By changing the namespace URI and the local name in the qname 
dc:creator, any property that can be serilized in RDF/XML can be 
included in this proposal.

There are some limitations but I believe they will be sorted in due course.

Jeremy





> 
> Although I appreciate that a primary goal of XHTML is human readability, I
> don’t necessarily think that this need follow for the application of
> ‘metadata’ in XHTML. Hence I consider the addition of a fixed set of
> additional tag properties to be somewhat restrictive. Instead – as I have
> already briefly discussed with Danbri – I favour a more extensible scheme,
> more closely related to cascading style sheets.  Using this approach
> multiple metadata properties could be grouped and reused for fragment, or
> document, classification and enhancement.
> 
> As the term ‘style’ seems somewhat inappropriate when applied to metadata,
> I currently prefer to think about collections of metadata for such purposes
> as ‘Aspects’. In the same manner as current CSS standards, Aspects could
> either be implemented as in-line sections within an XHTML document or
> referenced externally. Where XHTML has a specific requirement for the
> application of metadata, e.g. <link rel=”index” resource=”../index.htm”/>,
> reservation of a specific XHTML metadata namespace would allow browsers to
> interpret qualified properties in exactly the same way as fixed tags within
> the core language, e.g.
> 
> <XHTMLMeta:rel>
>       <XHTMLMeta:reltype>index<XHTMLMeta:reltype>
>       <XHTMLMeta:rawresource>../index.htm<XHTMLMeta:rawresource>
> <XHTMLMeta:rel>
> 
> I am currently experimenting with this type of approach to allow
> ‘functional styling’ and resource aggregation using descriminant
> Description Identification (apologies Danbri, if I have incorrectly quoted
> the latest terminology for ‘smushing’!)
> 
> 
> Phil Tetlow
> Senior Consultant
> IBM Business Consulting Services
> Mobile. (+44) 7740 923328
Received on Monday, 4 October 2004 11:11:20 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:17:13 GMT