Re: consumer-producer/pixel-perfection/*the* GRDDL result

An afterthought ...

A GRDDL transform designer striving after triple-perfection, such as the 
Amy in David's use case, is likely to also be concerned about what the 
consumer understands from these triples. For example, if the GRDDL 
Result is in fact an OWL DL graph, maybe they want the consumer to 
perform complete OWL DL reasoning, as it was designed for.

We could decide to redesign GRDDL to allow the publisher to put such a 
constraint on the consumer.

That would be a mistake.

The axis of variability of concern there (what reasoning to perform on 
an RDF graph) has been explored in depth by other groups, and the 
current situation works, despite this variability.

David's proposed designs are equally mistaken.

The axis of variability of concern in these (what preprocessing is 
performed when reading an XML document) has been explored in depth by 
other groups, and the current situation works, despite this variability.

Jeremy

-- 
Hewlett-Packard Limited
registered Office: Cain Road, Bracknell, Berks RG12 1HN
Registered No: 690597 England

Received on Friday, 22 June 2007 12:01:03 UTC