W3C home > Mailing lists > Public > public-prov-wg@w3.org > September 2011

Re: Roles

From: Satya Sahoo <satya.sahoo@case.edu>
Date: Fri, 16 Sep 2011 11:17:20 -0400
Message-ID: <CAOMwk6y9YCqsx7D8f9c9iPedGe5jXmn+h_bTDnKVaFkNRbeuVw@mail.gmail.com>
To: Paolo Missier <Paolo.Missier@ncl.ac.uk>
Cc: public-prov-wg@w3.org
Hi Paolo and Luc,

>Similarly:
>wasGeneratedBy(e1,pe1,qualifier(port="p1", order=1),t1)
>would become:
>wasGeneratedBy(e1 WITH {port="p1", order=1}, pe1,t1)
>this however assumes that one can always decide which side of the
wasGeneratedBy relationship the qualifiers belong to. It can be one or the
other, or both. >how would map to OWL?

>the qualifier {port="p1", order=1} is "linked" with the entity, it's also
relevant to the pe ... after all, emitting data on port p1.

Entity e1 was generated by PE pe1 on port p1 and its order is 1, in
pseudo-RDF

e1 wasGeneratedBy pe1
e1 wasGeneratedOnPort p1
e1 hasOrderOfGeneration 1

The above three assertions together with an application-specific inference
rule (since the relevance of order information is not common across all
application domains):
If Entity wasGeneratedBy PE AND Entity hasOrderOfGeneration n
then Entity is(nth)GeneratedEntityBy PE

Similar inference rule for port will capture the information that e1 is "an
entity emitted on port1" by pe1.

Does this capture the above scenario?

It is important to note that above pseudo-RDF does not use Roles - an
alternate modeling approach can use two possible roles - "entities generated
on port1" (generated by pe1 or pe2 etc.) and "entities whose order of
generation is 1" (generated by pe1 or pe2 etc.) and e1 assumes these roles.

> if the qualifiers logically "belong" with both sides, then you can't play
this game of "absorbing" them on one side, can you.  > They logically belong
where the data and the process meet, i.e., in neither "per se".
I am not sure I understand the above description.

Thanks.

Best,
Satya


On Thu, Sep 15, 2011 at 6:01 PM, Paolo Missier <Paolo.Missier@ncl.ac.uk>wrote:

>  Hi Luc,
>
> yes, and that was indeed in part my point: if the qualifiers logically
> "belong" with both sides, then you can't play this game of "absorbing" them
> on one side, can you.  They logically belong where the data and the process
> meet, i.e., in neither "per se". There are bound to be cases like this.
>
> -Paolo
>
>
> On 9/15/11 6:18 PM, Luc Moreau wrote:
>
>
> Hi Paolo,
>
> I don't understand why, in your example,
>
> wasGeneratedBy(e1 WITH {port="p1", order=1}, pe1,t1)
>
> the qualifier {port="p1", order=1} is "linked" with the entity,
> it's also relevant to the pe ... after all, emitting data on port p1.
>
> Luc
>
>
>
Received on Friday, 16 September 2011 15:18:04 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 13:06:41 GMT