Re: PROV-ISSUE-142 (Tlebo): Can roles only be Literals? [Data Model]

On Nov 6, 2011, at 12:01 PM, Luc Moreau wrote:

> Hi Paul,
> 
> I am not in favour of this,

:-(


> as explained in the previous email.

You said:

"It is intentionally that we use typed literals here, and not identifiers.

If we allow identifiers, then in effect we would have introduced a new
relation;  it would be better defined as prov-dm relation, rather
than hidden in a qualifier."


I'm afraid that I don't understand the differences between DM's literals, identifiers, relations, and qualifiers well enough.

If by "relation" you mean Usage, Control, Participation, and Generation, then YES I agree adding role to this would be concerning.

I like how it is currently stated:
"Qualifiers can be associated to relations, namely use and wasGeneratedBy, in order to further characterize their nature. Role is a pre-defined qualifier."

which lets me describe the relation by specifying the role that the Entity played.

I just don't want to tie our hands with a prov:role range of rdfs:Literal - we should be able to dereference the values of prov:role to find out more about it through Linked Data.

-Tim




> 
> What this shows, is that this WG can't reach consensus
> on role, and we might as well drop it from the model,
> because it uses too much bandwidth!


I just want a literal to be a Resource. Them I'm happy.
If we drop role, PROV-DM will be MUCH less useful.

-Tim




> 
> We should seriously consider this proposal!
> 
> Luc
> 
> On 06/11/11 15:46, Paul Groth wrote:
>> Hi Jim
>> 
>> I think to address Tim's issue we should just allow qualifiers to be Literals or Identifiers as defined by prov-dm.
>> 
>> Then one can use uris for roles without breaking compatibility with prov-DM
>> 
>> Was that your suggestion ?
>> 
>> Cheers
>> Paul
>> 
>> On Nov 6, 2011, at 16:19, Jim McCusker<mccusj@rpi.edu>  wrote:
>> 
>>   
>>> Make Roles resources like Entities. Classes of Roles (Creator,
>>> Publisher, PrincipleInvestigator) are instantiated for each Entity. Of
>>> course, this is rather similar to what's been rejected (?), but is
>>> still the best choice, IMO. This would be, for instance, "Barack
>>> Obama's role as president", as opposed to "the role of president",
>>> which would be a class.
>>> 
>>> Role could also be an extension of skos:Concept and allow you to
>>> express "the role of president" directly without custom instantiation.
>>> 
>>> Jim
>>> 
>>> On Sun, Nov 6, 2011 at 8:38 AM, Luc Moreau<L.Moreau@ecs.soton.ac.uk>  wrote:
>>>     
>>>> Ok, so, what's alternative suggestion ?
>>>> 
>>>> Professor Luc Moreau
>>>> Electronics and Computer Science
>>>> University of Southampton
>>>> Southampton SO17 1BJ
>>>> United Kingdom
>>>> 
>>>> 
>>>> On 6 Nov 2011, at 12:59, "Jim McCusker"<mccusj@rpi.edu>  wrote:
>>>> 
>>>>       
>>>>> This is a misunderstanding of a URI literal versus URI resource. When
>>>>> a URI resource is used, it can link to that resource when it has
>>>>> assertions made about it. This is not possible or intended with URI
>>>>> literals.
>>>>> 
>>>>> Jim
>>>>> 
>>>>> On Sun, Nov 6, 2011 at 2:36 AM, Luc Moreau<L.Moreau@ecs.soton.ac.uk>  wrote:
>>>>>         
>>>>>> Hi Tim,
>>>>>> 
>>>>>> But doesn't this include URIs by means of typed literals?
>>>>>> 
>>>>>> Professor Luc Moreau
>>>>>> Electronics and Computer Science
>>>>>> University of Southampton
>>>>>> Southampton SO17 1BJ
>>>>>> United Kingdom
>>>>>> 
>>>>>> On 6 Nov 2011, at 01:20, "Provenance Working Group Issue Tracker"<sysbot+tracker@w3.org>  wrote:
>>>>>> 
>>>>>>           
>>>>>>> PROV-ISSUE-142 (Tlebo): Can roles only be Literals? [Data Model]
>>>>>>> 
>>>>>>> http://www.w3.org/2011/prov/track/issues/142
>>>>>>> 
>>>>>>> Raised by: Timothy Lebo
>>>>>>> On product: Data Model
>>>>>>> 
>>>>>>> prov-dm, 5.5.1 Qualifier:
>>>>>>> 
>>>>>>> "The value associated with a role attribute must be conformant with Literal."
>>>>>>> 
>>>>>>> 
>>>>>>> 
>>>>>>> Will this prevent PROV-O from using URIs to cite roles?
>>>>>>> 
>>>>>>> Restricting roles to literals will be severely limiting for PROV-O and semantic web applications, since literals cannot be described or served as linked data, and thus consumers will be unable to determine more information about what the role means.
>>>>>>> 
>>>>>>> 
>>>>>>> 
>>>>>>>             
>>>>>> 
>>>>>> 
>>>>>>           
>>>>> 
>>>>> 
>>>>> --
>>>>> Jim McCusker
>>>>> Programmer Analyst
>>>>> Krauthammer Lab, Pathology Informatics
>>>>> Yale School of Medicine
>>>>> james.mccusker@yale.edu | (203) 785-6330
>>>>> http://krauthammerlab.med.yale.edu
>>>>> 
>>>>> PhD Student
>>>>> Tetherless World Constellation
>>>>> Rensselaer Polytechnic Institute
>>>>> mccusj@cs.rpi.edu
>>>>> http://tw.rpi.edu
>>>>>         
>>>> 
>>>>       
>>> 
>>> 
>>> -- 
>>> Jim McCusker
>>> Programmer Analyst
>>> Krauthammer Lab, Pathology Informatics
>>> Yale School of Medicine
>>> james.mccusker@yale.edu | (203) 785-6330
>>> http://krauthammerlab.med.yale.edu
>>> 
>>> PhD Student
>>> Tetherless World Constellation
>>> Rensselaer Polytechnic Institute
>>> mccusj@cs.rpi.edu
>>> http://tw.rpi.edu
>>> 
>>>     
> 
> 

Received on Sunday, 6 November 2011 21:45:32 UTC