Towards a precise, reliable definition of "information resource"

Leaving all issues about the necessity, utility, and
significance of defining a class of "information resources"
aside, I recommend to the TAG that it adopt a similar
approach as employed by RDF Core WG regarding the use of 
test cases to facilitate capturing concensus, by constructing
two lists: one that contains examples of resources that are 
agreed to be information resources, and one that contains
examples of resources that are agreed not to be information
resources.

I think a total of 20-25 resources per list should be sufficient.

The definition of the class "information resource" would
then be expressed/refined so that all agree that it reflects 
precisely the partitioning of those two lists.

Those two lists could (or should) be included in whatever
document officially presents that definition (not necessarily,
even if likely, AWWW) to reinforce the meaning of the
definition.

Furthermore, each resource in those two lists could be linked
to discussion about why it is or is not an information resource.
 
Thus, if someone wonders if some resource is an information 
resource, and they are not sure based on the text of the
definition, they can look at those two lists and see which
resources in those two lists most closely correspond to the 
resource in question and perhaps deduce a reasonably reliable 
answer to their question by such a comparison.

Regards,

Patrick


--

Patrick Stickler
Nokia, Finland
patrick.stickler@nokia.com
 

Received on Thursday, 21 October 2004 12:48:50 UTC