Re: PROV-ISSUE-56 (derivation-definition-does-not-imply-transitivity): Derivation as defined is not transitive [Conceptual Model]

In view of other responses to this question, and further thinking about 
use-cases, I'm shifting my position on this particular issue.  I see there *is* 
value in derivation being transitive.

The particular use-case I'm thinking about, which is derived from a real 
requirement in genetic factors and epidemiology, is this:  what do we need to 
know provenance-wise in order to repeat an analysis and get the same answer? 
Roughly, I think the answer is "same inputs, same processes".  And I think that 
to determine if the same data is available, a transitive closure of the inverse 
of derivation of the result is needed.  (Questions remain about the completeness 
of available information, but that's a separate issue.)

#g
--

Graham Klyne wrote:
> Short answer: no.
> 
> Longer answer:
> 
> I think we should focus on describing what needs to be described, and 
> allow the inferences to follow (or not) from that.  I think to construct 
> definitions to achieve desired inferences is putting the cart before the 
> horse.
> 
> Further, I think that there's a real danger that by focusing on 
> inferences rather than descriptions, we end up with terms whose 
> descriptive role is counter-intuitive, and which will, in the end, be 
> used or generated incorrectly by systems on the deployed web.
> 
> IMO, it's easier to add constraints later to enable inferences than it 
> is to work around unwanted constraints that are baked into a 
> vocabulary.  Particularly on an open-world monotonic logic language like 
> RDF (which I assume will provide the base language for actually 
> implementing these descriptions).
> 
> #g
> -- 
> 
> Luc Moreau wrote:
>> Hi all,
>>
>> Nice counter-example, Graham!
>>
>> We have the opportunity to define relationships with the properties we 
>> want them to have.
>>
>> Do we want (a form of ) derivation to be transitive?
>>
>> In the example that Graham provides, do you feel that A has some form 
>> of "influence" on C?
>> If so, would you like it to be automatically inferable in the 
>> provenance model?
>>
>> Regards,
>> Luc
>>
>>
>> On 07/29/2011 10:01 AM, Provenance Working Group Issue Tracker wrote:
>>> PROV-ISSUE-56 (derivation-definition-does-not-imply-transitivity): 
>>> Derivation as defined is not transitive [Conceptual Model]
>>>
>>> http://www.w3.org/2011/prov/track/issues/56
>>>
>>> Raised by: Graham Klyne
>>> On product: Conceptual Model
>>>
>>>
>>> [[ Given an assertion isDerivedFrom(B,A), one can infer that the use
>>> of characterized entity denoted by A precedes the generation of the
>>> characterized entity denoted by B.  ]]
>>> Where does this notion of "use" come from in the absence of some
>>> referenced activity?
>>>
>>> Concerning transitivity of derivation:
>>>
>>> Suppose:
>>> A has attributes a0, a1
>>> B having attributes b0, b1 is derived from A, with b0 being dependent 
>>> on a0
>>> C having attributes c0, c1, is derived from B with c1 being dependent 
>>> on b1
>>>
>>> So none of the attributes of C can be said to be directly or
>>> indirectly dependent on attributes of A, which by the given definition
>>> is a requirement for derivation of C from A.  Thus, as defined,
>>> derivation cannot be transitive.
>>>
>>> I don't really know if derivation should or should not be transitive,
>>> but the above seems to me like a problem of spurious
>>> over-specification.  My suggestion for now would be to focus on what
>>> really matters and see what logical properties fall out later.
>>>
>>>
>>>
>>>
>>>    
>>
> 
>

Received on Thursday, 4 August 2011 10:10:52 UTC