Re: proposal for an xlink data model

hello christopher.

> Thanks for posting this.  I looked it over - what does it add to the 
> infoset contributions described in "XML Linking and Style", <URL: 
> >?

first of all, i think that the data model should be a separate document, 
so that others (dom, css, ...) can reference it properly. more 
specifically, the infoset contributions in the older proposal have the 
following shortcomings:

- they do not allow resources in links which do not belong to an arc. 
xlink allows this, so it should be covered in the data model.

- arcs always have exactly one start resource and one ending resource. 
in xlink, an arc can have any number of start an ending points, so this 
should be covered in the data model.

- there is no way to point to a link from the markup that specifies that 

- i don't like the concepts of defining the model using "element 
information items". this in part is a result of the rather vague way of 
how the infoset's set of data types has been defined, but i think the 
xlink infoset items should be defined without reusing existing items.

- i found the modeling of the arcs a bit weird. arc items point to 
participant items which point to a resource, but the resources (if they 
are local resources) directly point to the arcs.

i hope this helps. i think that the link2rdf note also is a good 
starting point. i simply did not at it for a while because of some of 
its limitations which make it more restricted than xlink is.


erik wilde  -  tel:+41-1-6325132  -  fax:+41-1-6321035
           computer engineering and networks laboratory
           swiss federal institute of technology  (eth)
           * try not. do, or do not. there is no try. *

Received on Wednesday, 14 August 2002 04:21:14 UTC