PROV-ISSUE-54 (which-provenance-is-expected-to-be-retrieved?): What is it we expect to obtain when we say "retrieving the provenance of something" [Accessing and Querying Provenance]

PROV-ISSUE-54 (which-provenance-is-expected-to-be-retrieved?): What is it we expect to obtain when we say "retrieving the provenance of something" [Accessing and Querying Provenance]

http://www.w3.org/2011/prov/track/issues/54

Raised by: Luc Moreau
On product: Accessing and Querying Provenance

This issue is probably at the interface between the PAQ document and the Model document.
<br/>
But what are expecting to find when dereferencing a provenance-uri?  Syntactically, a set of assertions, as defined in the model, packaged in an appropriate manner. But what should those assertions relate to?
<br/>

-The past history of a given BOB? 
-The known uses of a given BOB?  (i.e. its future)
-A combination of both?
-Something else?
<br/>

If we dereference a provenance-uri twice, do we always get the same contents?  If not, then the provenance-uri points to a stateful resource.  If yes, then it seems that the provenance-uri is a cool URL.  Of course provenance may change over time (as we get to know more about things). Is it realistic to put a burden on the provider to honour all those cool URLs?
<br/>

Alternatively, if we perform an HTTP HEAD request on a static resource, and we find a provenance-uri in the header, is it always the same provenance-uri?


<br/>


Also, once this provenance-uri is minted, what is its time to live?  How long should the provider be able to serve these uris?

<br/>

To be inter-operable, I think that some clarification on these issues is required.

Received on Friday, 29 July 2011 08:00:19 UTC