RE: complementOf -> viewOf: proposed text

I think the issue arises not just between observers, but any time we talk about more than one kind of process. If I describe someone sitting down and eating, the hungryperson (who goes away when the eating process occurs) and sitting person (who appears during sitting) are alternates, but only in the interval between those two events. (I still need to give you a reason why the state of hungryperson and sitting person matter to fully justify the need for those specializations to be asserted - perhaps I want to know what affected the person's conversations with people - were they grouchy from hunger or not making eye contact or both and how did that affect the outcomes...). I think as soon as you have more than one process/are affecting more than one attribute of the entities, and care about the fixed state of those specializations, you can't avoid the issue...


> -----Original Message-----
> From: Paul Groth [] On Behalf Of Paul Groth
> Sent: Wednesday, January 18, 2012 8:35 AM
> To: Paolo Missier
> Cc: Myers, Jim; Paolo Missier; Stian Soiland-Reyes; Luc Moreau; public-prov-
> Subject: Re: complementOf -> viewOf: proposed text
> Hi All,
> I think the best way to look at this is if we see if alternateOf works from one
> perspective/observer. I think in this case it actually does even with transitivity.
> It just depends on what the single observer is trying to say about the situation.
> regards,
> Paul
> Paolo Missier wrote:
> > Jim
> >
> > (BTW I can't keep up either, esp. with a pile of students exams to mark...).
> >
> >    On 1/17/12 5:29 PM, Myers, Jim wrote:
> >> I can see how the definition you have leads to the consequence you state,
> but it seems like the use case here is one we should be able to support -
> someone reports on the activities of the customer-in-the-red-chair over time
> and others report that Paolo and Stian were in the chair at various times and
> we'd like to have enough prov information to allow users to figure out who
> did what.
> > critically (I think), you have reintroduced time which is not in
> > Stian's example. If we discount time as Stian and I did, i.e. we
> > assume two observers look at the same scene at the same time, would my
> > conclusion be more reasonable?  If so, then maybe the old idea of
> > accounting for "temporal overlaps" in the definition of alternate needs to
> come back. And if it does, then we are back to some earlier "pseudo-
> transitivity" property that was discussed a while ago, and was dismissed as
> awkward, and I would agree on giving up on transitivity altogether. Would that
> help address your other question:
> >> If alternateOf is not capable of doing this, do we have some other
> mechanism that can? Or is the use case out of scope?
> > --Paolo
> >
> >> -- Jim
> >> ________________________________________
> >> From: Paolo Missier []
> >> Sent: Tuesday, January 17, 2012 12:16 PM
> >> To: Myers, Jim
> >> Cc: Paolo Missier; Stian Soiland-Reyes; Luc Moreau; public-prov-
> >> Subject: Re: complementOf ->   viewOf: proposed text
> >>
> >> Jim
> >>
> >> (haven't read your latest email, except to spot that you disagree
> >> with transitivity, but I am not prepared to argue just now)
> >>
> >> I think unless you are prepared to accept that they are different
> >> characterization of the same real-world thing, then they should not be
> alternates of each other. (at least) one of the two is not the customer in the
> red chair.
> >>
> >> -Paolo
> >
> >
> >
> >
> >

Received on Wednesday, 18 January 2012 14:11:37 UTC