Re: reviewer feedback on prov-o ontology

Hi James,

I would have thought that what can be expressed in prov-dm is what is 
By that, I mean, that we can express it in various representations: rdf, 
xml, etc, without
losing anything.

Each mapping of prov-dm to a given technology will come with its nice 
The one being discussed in this thread, with the owl ontology, is about 
some structure to the various prov-dm concepts and allow for the kind of 
expressivity Jim
described in his email.
I have seen OO mapping that introduce abstract classes, also nice in 
that context.

But to me, to be interoperable, it's important that we distinguish the 
core constructs
(which can be mapped to prov-dm, and other representations) from the 
mapping specific

Never we claimed that we would define all concepts related to 
provenance. Only a core that we
can exchange in an interoperable manner, so that systems, making use of 
their respective
technologies can still make sense of it.


On 02/23/2012 02:51 PM, James Cheney wrote:
> Hi Luc,
> Here and in other threads I'm not sure what you mean by non-aligned. 
>  As discussed concerning ISSUE-253, there are lots of other things 
> that can be said using the RDF vocabulary that don't necessarily 
> correspond to things that can be said in DM syntax.  But so what?  How 
> does this harm interoperability?  What property do you expect that 
> isn't established here (e.g. some kind of roundtripping property).
> I think it is interesting to make this explicit so that we can try to 
> achieve it (or agree whether it's achievable/desirable).
> --James
> On Feb 23, 2012, at 2:41 PM, Luc Moreau wrote:
>> Hi Jim,
>> I didn't say it was a bug. I said it was not aligned.
>> It's a feature, maybe, but that's not expressible in prov-dm, which means
>> that other implementations e.g. java, xml, or whatever would not 
>> understand that feature.
>> So, from an ontological viewpoint, a nice feature, but one that does 
>> not help with
>> interoperability.
>> Luc
>> On 02/23/2012 02:21 PM, Jim McCusker wrote:
>>> On Thu, Feb 23, 2012 at 9:30 AM, Luc Moreau 
>>> < <>> wrote:
>>>     Furthermore, the ontology allows for instances of involvements
>>>     to be expressed, without
>>>     specifying its subclass (Usage, Generation, etc). This is not
>>>     aligned with the data model.
>>> This is a feature, not a bug. Even if Involvement were defined as 
>>> equivalent to the union of subclasses, it would still be possible 
>>> (and consistent) to assert that something is an Involvement without 
>>> saying what the subclass is. We simply wouldn't know.
>>> Jim
>>> -- 
>>> Jim McCusker
>>> Programmer Analyst
>>> Krauthammer Lab, Pathology Informatics
>>> Yale School of Medicine
>>> <> | (203) 
>>> 785-6330
>>> <>
>>> PhD Student
>>> Tetherless World Constellation
>>> Rensselaer Polytechnic Institute
>>> <>
>>> <>
>> -- 
>> Professor Luc Moreau
>> Electronics and Computer Science   tel:   +44 23 8059 4487
>> University of Southampton          fax:   +44 23 8059 2865
>> Southampton SO17 1BJ     
>> United Kingdom
> The University of Edinburgh is a charitable body, registered in
> Scotland, with registration number SC005336.

Professor Luc Moreau
Electronics and Computer Science   tel:   +44 23 8059 4487
University of Southampton          fax:   +44 23 8059 2865
Southampton SO17 1BJ               email:
United Kingdom           

Received on Thursday, 23 February 2012 15:20:12 UTC