Norman Walsh wrote:

>  In conversations today on the XML Linking WG, prompted by Erik Wilde's
>  proposal for an XLink Data Model[1], I was reminded of another
>  possible solution. [...]
> 
>  This other alternative answer is to use infoset augmentation. By
>  modeling linking semantics in terms of additions to the infoset, we
>  can make a clear separation between the syntactic constructs used to
>  identify links and their meaning.

I think this is an excellent idea.

>  So, for example, an HTML browser can construct appropriate link
>  information items from HREF, SRC, and LONGDESC attributes in the XHTML
>  by appealing to its own understanding of the XHTML vocabulary.
>  Similarly, an XLink-aware application can construct them from XLink
>  1.0 attributes.

>  As an added bonus, this method allows an application that understands
>  links to treat both simple and extended links uniformly. It operates
>  on the linking data model without regard to how that was constructed,
>  whether the links expressed came originally from simple links in the
>  source document or a set of external linkbases.

... or a CSS-like "linking stylesheet," or an AF-like processor,
or a validator, or any number of other pre-processing transformations.
This would allow a lot of flexibility in how XLink could be
applied.

>  On the other hand, this is an, uhm, LSI, and is sure to be greeted by
>  some with the same enthusiasm as the PSVI :-).


AFAICT, it's not so much the idea of an augmented infoset
that people find disagreeable about the PSVI, it's the
nature (and quantity!) of the data that a W3C XML Schema
validator adds to the Infoset.

Come to think of it, if the Infoset augmentation is expressed
as a set of attributes with a distinguished namespace name,
instead of as a new set of [information item]s, then
the current XLink draft could work unchanged.  You just
need to explicitly state that the xlink:* attributes need
not be present in the source document, and may instead
be added by some (unspecified) preprocessing transformation.


--Joe English

  jenglish@[...].com

-----------------------------------------------------------------
The xml-dev list is sponsored by XML.org <http://www.xml.org> , an
initiative of OASIS <http://www.oasis-open.org> 

The list archives are at http://lists.xml.org/archives/xml-dev/

To subscribe or unsubscribe from this list use the subscription
manager: <http://lists.xml.org/ob/adm.pl>