Re: PROV-ISSUE-4: Defining Agent using FOAF's definition

Folks -

[Cut-and-paste from another email I sent that I don't see coming through for 
some reason]

I suggest that there is a distinction made between "System Agent" and "Human 
Agent" in http://www.w3.org/2011/prov/wiki/ConceptAgent since this may be a 
prevalent case.  I'm aware of UML extensions, for example, that this distinction 
was made and was very useful.  In the case of provenance, there may be automated 
programs that are linked to  processes per description of the model insofar 
without direct human interaction.  This can be quite crucial as the trust models 
for system agent versus human agents can be quite different and provenance 
applications need strong trust models (to our experience from a vendor's 
perspective) Reference examples -

          * http://www.google.com/url?sa=t&source=web&cd=1&ved=0CBkQFjAA&url=http%3A%2F%2Fwww.azouk.com%2F179447%2Fno-title%2Fopen&rct=j&q=UML%20Nunes%20extension%20%22System%20Actor%22&ei=UyoJTofGL-ja0QGqtKx0&usg=AFQjCNEf0TNStqdsjtdrnM7DYEV-V4qbIg&cad=rja
          * http://books.google.com/books?id=Vl1H266pOVEC&pg=PA385&lpg=PA385&dq=UML+Nunes+extension&source=bl&ots=RDQgUOQ5VV&sig=2Ac8UbIhexB3a80DR8QN0DW2slY&hl=en&ei=PioJTvzEFKnt0gGX9oiiAQ&sa=X&oi=book_result&ct=result&resnum=1&ved=0CBYQ6AEwAA

Best.

On 6/24/11 10:01 AM, Myers, Jim wrote:
>> Agreed. :-)  Although it would also be part of establishing trust to also see
>> what was NOT done (eg. "Sterilize the microscope plate") - but knowledge of
>> what 'should have been done' is clearly out of scope.
>> Considering what DID happen would typically be compared to the known
>> practice/procedure/workflow/etc - and our provenance model would have
>> enough information to answer queries to validate that side of the story.
> And to establish the 'recipe link' from the provenance to the procedure...