Re: quick comment on Note in ProvRDF mapping

Hi James,

Response interleaved.

On 02/14/2012 10:29 AM, James Cheney wrote:
> What's the use case?  More generally, in WD3 at least, there are no examples of alternateOf or specializationOf in use (with or without attributes).
>
>    

You will recall that this is the first draft of this section with 
alternateOf and specializationOf ...

I think that for prov-dm, we are coming to the conclusion that we will 
not define whether these
relations are symmetric/transitive etc.
Some communities may want to define specialized version that are 
symmetric/transitive.


> If we want to make things really uniform, we could identify a common "template" for all of the record forms
>
> record(id, blah, blah, ... , attrs)
>
>    

Essentially, that's the template we followed, with blah, blah varying 
for each record type.

> Most of the rules in ProvRDF have this form already.
>
>    

Ultimately, it's a design decision for the WG.

1. Do we adopt a common pattern for all elements and relations?
2. Do we 'customize' relations according to use cases we have encountered
        (which would mean, potentially, dropping attributes/ids for 
hasAnnotation/alternatedOf)

I can see pros and cons for a common pattern approach:
pluses:
  - simpler to understand for the user, no special case
  - systematic handling of mapping to rdf/xml/etc
  - attributes are the key mechanism for extensibility, and hence it's 
good to have them

minuses:
  - not obvious (now!) how attributes can be used in some cases
  - may force the introduction of an OWL class, which wouldn't be 
required otherwise


Luc
> --James
>
> On Feb 14, 2012, at 9:33 AM, Luc Moreau wrote:
>
>    
>> Hi James,
>>
>> I think it was an oversight on our behalf (Paolo and I) not to include
>> an id for alternateOf/specializationOf. In our working copy,
>> http://dvcs.w3.org/hg/prov/raw-file/default/model/working-copy/towards-wd4.html
>> we have added them.
>>
>> I also take the view that if we have an id then we have attributes, and vice-versa.
>>
>> As a minimum, subtyping would be useful for these relations.
>> You will also recall, very early discussions about mapping of attributes for IVPof.
>> This could also be encoded with attributes.
>>
>> Cheers,
>> Luc
>>
>>
>> On 02/14/2012 09:18 AM, James Cheney wrote:
>>      
>>> While we're on the subject, I'm no sure why alternateOf and specializationOf have attributes now, other than uniformity.
>>>
>>> I think that if the relation has an id describing the relationship (used/Usage, rtc.) Then attributes make sense.  If an id doesn't make sense then attributes don't either - in RDF we need an id to hang the attributess off of.
>>>
>>> I think that brevity should take precedence over uniformity, else we'll reinvent RDF or XML.
>>>
>>> --James
>>>
>>>
>>>
>>>        
>> -- 
>> Professor Luc Moreau
>> Electronics and Computer Science   tel:   +44 23 8059 4487
>> University of Southampton          fax:   +44 23 8059 2865
>> Southampton SO17 1BJ               email: l.moreau@ecs.soton.ac.uk
>> United Kingdom                     http://www.ecs.soton.ac.uk/~lavm
>>
>>
>>
>>      
>
>    

-- 
Professor Luc Moreau
Electronics and Computer Science   tel:   +44 23 8059 4487
University of Southampton          fax:   +44 23 8059 2865
Southampton SO17 1BJ               email: l.moreau@ecs.soton.ac.uk
United Kingdom                     http://www.ecs.soton.ac.uk/~lavm

Received on Tuesday, 14 February 2012 10:53:54 UTC