W3C home > Mailing lists > Public > public-openannotation@w3.org > November 2012

Re: F2F Decision: Provenance

From: Graham Klyne <Graham.Klyne@zoo.ox.ac.uk>
Date: Fri, 02 Nov 2012 16:25:13 +0000
Message-ID: <5093F3E9.5060608@zoo.ox.ac.uk>
To: Robert Sanderson <azaroth42@gmail.com>
CC: Paolo Ciccarese <paolo.ciccarese@gmail.com>, public-openannotation <public-openannotation@w3.org>
On 02/11/2012 15:24, Robert Sanderson wrote:
> I agree that prov:alternativeOf is slightly broader than our use case for
> oa:equivalent, but I'm not (yet) convinced that we need to revert to our
> own specialization.

A fair point I think.  Maybe (see below).  (flip flop? what, me? :) )

> Is there a situation in which someone would want to use prov:alternateOf
> and it would NOT be consistent with what we want to use it for? If not,
> then I would definitely prefer to keep it.  If yes, then let's create a
> specialization.

That's a reasonable question.  But if the answer is yes, I don't think a 
subproperty works either.

Another is:  is there a situation in which one would want to distinguish between 
the broader meaning of prov:alternativeOf and the tighter meaning of 
oa:equivalent.  E.g., there might be useful inferences from the latter that are 
not supported by the former.  If yes, a subproperty might be desirable.

> The semantics, in my opinion, are derived from the decision to conflate
> Annotation as a concept and the Document that encodes it.  Thus two
> annotation documents are the equivalent if they encode the same conceptual
> annotation, potentially with different metadata and necessarily with a
> different URI.

Hmmm... sounds to me a bit like a lack of clarity over use and mention?  It 
sounds as if oa:equivalent is a relation between *mentions* of some common 
assertion.  I'm not sure that properly qualifies as a prov:alternativeOf.  I'm 
not sure about this, but I'm having a hard time seeing a mention of some 
assertion being a constrained form of something that is common to all such mentions.

(If an infinite number of monkeys were to produce a script for Hamlet, would 
that be a specialization of something in common with Shakespear's version of same?)


> On Fri, Nov 2, 2012 at 7:33 AM, Paolo Ciccarese
> <paolo.ciccarese@gmail.com>wrote:
>> Dear Graham,
>> thank you for your feedback and welcome to the group!
>> I am assuming
>> a prov:alternativeOf v
>> was meant to be
>> a prov:alternativeOf b
>> The current definition of oa:equivalent is: The subject and object
>> resources of the oa:equivalent relationship represent the same Annotation,
>> but potentially have different metadata such as generator, generated and
>> serialization format. oa:equivalent is a symmetrical relationship; if A
>> oa:equivalent B, then it is also true that B oa:equivalent A.
>> Basically it is a mechanism to allow multiple 'copies' of the same
>> annotation. Each copy identified by a different URI and can have different
>> metadata.
>> Given this and given your explanation, I believe prov:alternativeOf has a
>> broader meaning than oa:equvalentTo. Maybe we can keep our original
>> property and declare it a sub-property of prove:alternativeOf?
>> We would appreciate more feedback from the prov group on this matter.
>> Best,
>> Paolo
>> ps: we will fix prov:generatedAt
>> On Thu, Nov 1, 2012 at 6:54 PM, Graham Klyne <graham.klyne@zoo.ox.ac.uk>wrote:
>>> Hi all,
>>> I may be coming late to this party.  Sorry if I grab the wrong end of the
>>> stick here.  I'm a participant in the prov group, but here am speaking
>>> strictly for myself - other members of the group may disagree.
>>> ...
>>> prov:alternativeOf isn't quite the same as "equivalence", though it's
>>> possible that it's similar to ao:equivalent.  I don't know if ao:equivalent
>>> means more or less than one might expect of "equivalent".
>>> My interpetation of prov:alternativeOf is roughtly
>>> exists(c)
>>> a prov:specializationOf c
>>> b prov:specializationOf c
>>> |-
>>> a prov:alternativeOf v
>>> where
>>> a prov:specializationOf c means that the resource a is resource c
>>> constrained to some interval or context or situation.
>>> E.g. (Boston in 1776) prov:specializationOf (Boston)
>>> Following this, we might have
>>> (Boston in 1776) prov:alternativeOf (Boston in 2012)
>>> Are they equivalent?  I'd say not.  But do they in some sense refer to
>>> the same thing?  I'd say so.
>>> If that kind of semantics works for AO, then fine, but I'm suspecting it
>>> may be somewhat different to what you might expect.  It's not, for example,
>>> like the old rdf:Alternate class.
>>> ...
>>> As far as I'm aware, there is no prov:generatedAt property.  Do you mean
>>> prov:generartedAtTime?  (http://www.w3.org/TR/prov-o/#**generatedAtTime<http://www.w3.org/TR/prov-o/#generatedAtTime>
>>> )
>>> #g
>>> --
>>> On 01/11/2012 10:39, Antoine Isaac wrote:
>>>> Hi all,
>>>> I guess the lack of reaction means everyone agrees :-)
>>>> I may have trouble the very idea of representing oa:Annotation as direct
>>>> result
>>>> of the generating, as opposed to the direct result of annotating. But
>>>> I'll
>>>> clearly need some more time to get my head around it.
>>>> One trivial for now is replacing oa:generator with oa:generatedBy.
>>>> This makes the property seem very close to prov:wasGeneratedBy, in a
>>>> context
>>>> where OA and PROV would be used together. While they are quite different
>>>> in
>>>> reality: range of oa:generatedBy would be agent, range of
>>>> prov:wasGeneratedBy is
>>>> prov:Activity.
>>>> Antoine
>>>>   This part of the discussion covered two primary topics related to
>>>>> provenance and the W3C Provenance Ontology.
>>>>> 1.  Can we replace oa:equivalent with something from the Prov work?
>>>>> Decision:  Yes, prov:alternateOf is semantically identical to
>>>>> oa:equivalent
>>>>> Thus we'll simply replace all mentions of oa:equivalent in the
>>>>> specification with prov:alternateOf
>>>>> 2.  What is the relationship between the current (simple) provenance
>>>>> information recorded for an annotation, and the Prov work?
>>>>> Decisions:
>>>>>    - Replace oa:generated with prov:generatedAt, as they are
>>>>> semantically identical
>>>>>    - Replace oa:generator with oa:generatedBy, and subclass of
>>>>> prov:wasAttributedTo
>>>>>    - Replace oa:annotated with oa:annotatedAt (to follow the generatedAt
>>>>> pattern)
>>>>>    - Replace oa:annotator with oa:annotatedBy
>>>>>    - Include a diagram of the mapping in the specification
>>>>> (Diagram attached)
>>>>> Thanks,
>>>>> Rob&  Paolo
Received on Friday, 2 November 2012 16:26:14 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:38:20 UTC