Re: PROV-ISSUE-466 (must-entities-invalidate): Must all entities invalidate? [prov-dm-constraints]

On Mon, Aug 6, 2012 at 5:15 PM, Luc Moreau <l.moreau@ecs.soton.ac.uk> wrote:
> In prov-dm, we say "Entities have a duration", and "An activity is something
> that occurs over a period of time".
> Those two inferences only reflect what we say in prov-dm.

A duration does not have to be finite.  (IMHO). Of course, practically
this would primarily only make sense for conceptual entities. However,
we encourage to use PROV also for provenance of conceptual entities
and activities.


> Furthermore, in activity(a,-,-,attrs) we said that the two time positions
> are expandable, i.e.
> there exists t1, t2, such that activity(a,t1,t2,attrs).  This didn't seem a
> problem to say there exists t2, so why is it
> a problem to inference a end event.

In DM, an activity end time is optional. I have previously queried as
to what 'optional' means, if it means it is implied or not specified.
DM does not specify this, but here in Prov-Constraint  we are equiring
that the time exists, and thus that the activity must end; and
similarly that all entities must be invalidated.


This is a stronger requirement that I think we need to agree on at a
WG level. I can see the 'beauty' or 'consistency' argument to have
invalidation match generation (and same for activity start/end), but
we have previously agreed that PROV would not be describing things
that will happen in the future, or things that would have happened if
something was different.   (Thus we don't provide any details for
plans or intended usages).

In a normalized PROV instance, every entity and activity will end.
This seems a bit odd, as perhaps those activities or entities will
never end. Arguing that everything must end is a philosophical
argument that I would rather we did not delve to deep into.


Note that I am not insisting to remove requirement for invalidation
and activity end time, I would agree on keeping them (perhaps with a
note) if the WG votes that this is OK. (I would vote 0 for the sake of
not blocking).


-- 
Stian Soiland-Reyes, myGrid team
School of Computer Science
The University of Manchester

Received on Wednesday, 8 August 2012 09:32:00 UTC