W3C home > Mailing lists > Public > public-prov-wg@w3.org > February 2012

Re: PROV-ISSUE-242 (TLebo): generated twice? [prov-dm]

From: Graham Klyne <graham.klyne@zoo.ox.ac.uk>
Date: Tue, 07 Feb 2012 15:33:47 +0000
Message-ID: <4F31445B.8070202@zoo.ox.ac.uk>
To: Provenance Working Group <public-prov-wg@w3.org>
I think this is an example of the kind of possible simplification we discussed 
in the F2F.

It would be syntactically valid, and for "scruffy" provenance it might even make 
some kind of sense (e.g. for successive posts on a given blog).

But (assuming the generation time shows up directly or indirectly as an 
attribute value in the model theory) it would not be satisfiable within the 
model theory - by which I mean there would be no interpretation allowed by the 
model theory that could make the combined assertions true.  This would be a 
reflection that for "proper" provenance, there are certain additional 
constraints to be satisfied if the provenance inference are to be validly applied.

In summary, I think it's in line with what we agreed that the initial DM 
description says nothing (or very little) about this, and that we leave it to 
the explanation of the more formal handling of provenance to introduce these ideas.


On 06/02/2012 21:45, Provenance Working Group Issue Tracker wrote:
> PROV-ISSUE-242 (TLebo): generated twice? [prov-dm]
> http://www.w3.org/2011/prov/track/issues/242
> Raised by: Timothy Lebo
> On product: prov-dm
> Is the following valid?
> wasGeneratedBy(e,,2001-10-26T21:32:52)
> wasGeneratedBy(e,,2011-11-27T21:23:34)
> If no, where does DM say so?
> Thanks,
> Tim
Received on Tuesday, 7 February 2012 15:50:49 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:58:12 UTC