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

Re: PROV-ISSUE-207 (start and end records): start and end records [prov-dm]

From: Simon Miles <simon.miles@kcl.ac.uk>
Date: Fri, 3 Feb 2012 12:21:48 +0000
Message-ID: <CAKc1nHfd8B1a9m=eKCACQJpNZgGR1GJdaiDpNpnTxqycb4LDUA@mail.gmail.com>
To: Provenance Working Group WG <public-prov-wg@w3.org>
To record what I was proposing today regarding events, somewhat copied from
what Tim was going to propose :)

1. Events are just instants of interest, used to help explanation but not
firstclass modelled concepts

2. Assertions about the start event, wasStartedAt(activity, time), are
separate from assertions about responsibility for starting the activity,
wasStartedBy(activity, agent)

3. wasStartedBy is still a subtype of wasAssociatedWith

4. For conformity, other relationships associated with events are treated
similarly, e.g. used(activity, entity) and usedAt(activity, time)

5. Atfributes of relationships describe the relationships, which includes
an implied event but is not the same as the event, i.e. some attributes may
not be instantaneous

6. PROV-O does not change for existing qualified relationships, but we
explicitly do not interpret qualifiedinvolvement as an event for reason 5
above, e.g.

:qi a prov:Usage
 prov:atTime "2012-02-03T11:35";
 prov:hadRole www.example.com/crimeFile#author.

would be interpreted as information about a usage relationship asserted
between an activity and entity. That relationship is associated with an
event and prov:atTime (or however modelled now) denotes the time of that
event. hadRole describes the relationship, not the event.


On Wednesday, 21 December 2011, Provenance Working Group Issue Tracker <
sysbot+tracker@w3.org> wrote:
> PROV-ISSUE-207 (start and end records): start and end records [prov-dm]
> http://www.w3.org/2011/prov/track/issues/207
> Raised by: Luc Moreau
> On product: prov-dm
> We have recently voted on definitions for generation and usage records.
> Start and end records are the corresponding 'instantaneous events' for
> We should therefore define start/end records as representation of
activity start/end events.
> Second, time is handled differently for entities and activities. The
placeholders for time information related to entities is in generation and
usage events. But for activities, it's in the activity record.
> Shouldn't activity start/end time be moved to start/end record?
> If we do move activity start/end time to start/end record, then shouldn't
we also make the agent optional, for cases where we want to assert time
without knowing the agent.
> This latter proposal may also have implications for the prov-o ontology.

Dr Simon Miles
Lecturer, Department of Informatics
Kings College London, WC2R 2LS, UK
+44 (0)20 7848 1166

Efficient Multi-Granularity Service Composition:
Received on Friday, 3 February 2012 12:22:27 UTC

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