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

Re: [provo] Difference between wasInformedBy and wasStartedByActivity (ttl examples)

From: Stian Soiland-Reyes <soiland-reyes@cs.manchester.ac.uk>
Date: Tue, 8 May 2012 17:06:58 +0100
Message-ID: <CAPRnXt=MUNH9S9PDa1w-Ba0YKALp_pFcttrqOzZCQjU4hUfJLA@mail.gmail.com>
To: Daniel Garijo <dgarijo@delicias.dia.fi.upm.es>
Cc: Timothy Lebo <lebot@rpi.edu>, public-prov-wg@w3.org, Jun Zhao <jun.zhao@zoo.ox.ac.uk>, Khalid Belhajjame <Khalid.Belhajjame@cs.man.ac.uk>
That dual-mode style would confuse the interpretation of wasStartedBy, the
activity becomes a token (and thus and entity)

Perhaps better would be to add the activity as an optional parameter to
wasStartedBy in dm and add prov:hadActivity/prov:startedByActivity to
prov:Start. It would mirror the activity of derivations.

An activity start, written wasStartedBy(id,a,e,t,a2, attrs) in PROV-N, has:

id: an OPTIONAL identifier for the activity start;
activity: an OPTIONAL activity (a2) which generated the (possibly
unspecified) entity (e)

attributes: an OPTIONAL set (attrs)of attribute-value pairs representing
additional information about this activity start.

Then mirror this for wasEndedBy.

Stian Soiland-Reyes, myGrid team
School of Computer Science
The University of Manchester
On May 8, 2012 3:03 PM, "Daniel Garijo" <dgarijo@delicias.dia.fi.upm.es>

> Hi Stian,
> instead of removing the constraint that entity and activity are disjoint
> we could
> also (as another possibility) have activities OR entities as possible
> domain
> of wasStartedBy. Now that we agreed on having an OWL-RL ++ profile,
> this would be possible.
> Thus, we would drop wasStartedByActivity, since wasStartedBy would
> cover already the desired functionality, right?
> Thanks,
> Daniel
> 2012/5/8 Stian Soiland-Reyes <soiland-reyes@cs.manchester.ac.uk>
>> On Mon, May 7, 2012 at 3:48 PM, Timothy Lebo <lebot@rpi.edu> wrote:
>> > +1, repositioning wasStartedByActivity as a "blurrier" form of
>> wasStartedBy seems to finally find a place for it in the model.
>> > Though, like Khalid, I'm not sure it will be used much, or correctly.
>> It will certainly still be confusing, as it was for me. As you said,
>> most wasStartedBy() would also come with a twin used() relationship
>> (and therefore imply a wasInformedBy() relation).   At some point
>> wasStartedBy was sub-property of wasInformedBy (making the choice
>> simple) - but not anymore.
>> As Luc raised, why not also wasEndedByActivity,  wasStartedByAgent etc.?
>> So it might just not be worth it to keep wasStartedByActivity(). It's
>> a bad sign if it's confusing to even the ontology designers, then how
>> is any meaningful provenance exchange happen, where one party apply
>> wasInformedBy like wasStartedByActivity, and the other the opposite?
>> A second solution would be to remove the constraint that activity and
>> entity are disjoint. Then you could say wasStartedBy(a2, a1),
>> wasEndedBy(a2, a3) etc. - the activity can play the role of an entity
>> as well, rather than inventing invisible phantom token entities. We
>> are talking blurry provenance here, right, we don't know quite the
>> nature of the interaction.
>> > How can it be reframed so that wasStartedByActivity can "grow" in
>> details like Derivation does with hadActivity, hadUsage, and hadGeneration?
>> By adding a separate wasStartedBy() I would believe you have given all
>> the information (as an activity can only be started once).  Or is it
>> allowed to be wasStartedBy() two or more entities..? Luc?
>> --
>> Stian Soiland-Reyes, myGrid team
>> School of Computer Science
>> The University of Manchester
Received on Tuesday, 8 May 2012 16:07:28 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:51:14 UTC