Mark, Ralph, and Ben meet up

Hi team,

Ralph, Mark, and I met up for a couple of hours today to chat about  
remaining issues, and we thought fairly deeply about the inheritance  
issues, which are not quite all resolved yet. Right now, the status  
seems to be as follows:

1) the ABOUT attribute is inherited from parent elements as far up  
the hierarchy as necessary

2) META and LINK are special, in that they are meant solely to attach  
metadata to their immediate parent element. In other words, when you  
see a META or LINK, here's how you determine its subject:
a) if the META or LINK has its own ABOUT attribute, then that's its  
subject
b) if the META or LINK does not have its own ABOUT attribute, but if  
its parent element has an ABOUT attribute, then that is the subject.
c) if the META or LINK does not have its own ABOUT attribute, and  
neither does its parent element, then the subject of the assertion is  
a bnode represented by the parent element itself.
d) one final special case: if the parent element of the META or LINK  
is itself a META or LINK, then the subject of the child META or LINK  
is the statement represented by the parent META or LINK. In other  
words, that's how you get reification.

3) in RDF/A there is no way to have a single URI serve as both  
subject and object in different assertions, except by literally  
copying the URI text. This is different from RDF/XML, where the  
object of an assertion can be packaged together with its own assertions.


Mark and I will work on specific examples that detail out the above  
principles. Please send comments, thoughts, or feel free to stay  
silent until the examples are available :)

-Ben

Received on Friday, 30 September 2005 01:12:53 UTC