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

RE: complementOf -> viewOf: proposed text

From: Myers, Jim <MYERSJ4@rpi.edu>
Date: Thu, 19 Jan 2012 13:29:51 +0000
To: Luc Moreau <L.Moreau@ecs.soton.ac.uk>
CC: Paolo Missier <Paolo.Missier@ncl.ac.uk>, Paolo Missier <paolo.missier@newcastle.ac.uk>, Stian Soiland-Reyes <soiland-reyes@cs.manchester.ac.uk>, "public-prov-wg@w3.org" <public-prov-wg@w3.org>
Message-ID: <3131E7DF4CD2D94287870F5A931EFC2302A534DE@EX14MB2.win.rpi.edu>
Tried to read ahead in the thread, but let me come back to my statement about entities representing multiple things:

Perhaps the issue is that we are trying to create a technical definition of prov:thing that is different than our intuitive sense. What I was trying to say is that we're going to see entities that represent multiple things in the normal English sense of the word because those things already have some characterization in them. A prov:thing is really meant to be the raw stuff - whatever's in the chair whether you consider that to be part of something mobile (a person) or immobile (the customer-in-the-chair). I think our model doesn't have to solve the question about what raw stuff is - that's for philosophers. The key for us is that when a process happens, to one entity that characterizes a prov:thing, we require that it happen to /effect all entities that exist during that process and characterize that same prov:thing. The activity that happens may create one entity and just be part of the lifecycle of another, but specialization and alternative let us know that there's a connection between the history of these two entities.

Alternate is so hard to define because we're are having to define entities in terms of the things that people have named, which are not prov:things - they are prov:things that have some characterization in terms of some theory of the world.  (When fully characterized with respect to what the asserter wants to report, we have entities.)

I think we're debating only two real definitions of alternative with some confusion:
Transitivity makes sense if two entities characterize the same prov:thing - paolo-in-the-chair and customer-in-chair-named-paolo for example. Transitivity does not work if we allow that relationship between two colloquial things that at some point characterize the same prov:thing - e.g. paolo and customer-in-chair. Pseudo-transitivity  - i.e. transitivity during the time when both entities characterize the same prov:thing would still work, but then we need to know the time.event bounds on the alternative relationship to calculate when pseudo-transitivity applies (i.e. to infer that some entity specializing paolo between the sitting and standing events must characterize the same prov:thing as a specialization of customer-in-chair over the same interval because 'poalo alternativeof customer-in-chair between sitting and standing'.

So two choices (could we vote?):

Alternative applies only between two entities that characterize the same prov:thing over the same interval and is transitive: to use it in the customer-in-chair case you have to create specializations of both poalo and customer in chair.

Alternative applies between entities that at some point in their lifecycles describe the same prov:thing. Transitivity cannot be applied. Having a time/event interval specified on the relationship would allow inference that there are specializations of both entities that are aliases (symmetric specializations of each other?) Pseudotransitivity in this way could be a separate vote...


 Jim

> -----Original Message-----
> From: Luc Moreau [mailto:L.Moreau@ecs.soton.ac.uk]
> Sent: Wednesday, January 18, 2012 9:48 AM
> To: Myers, Jim
> Cc: Paolo Missier; Paolo Missier; Stian Soiland-Reyes; public-prov-wg@w3.org
> Subject: Re: complementOf -> viewOf: proposed text
> 
> Hi Jim
> 
> A simple English sentence seems to question seriously the model:
> 
>  One entity may be a characterisation for multiple things.
> 
>  Entity( e1, [ colour = blue ])
> 
> Indeed this characterise many things including
>  Tshirt-of-Luc-in-Boston-2011
> And
>   Luc-new-car-2012
> 
> How would that work with characterisation intervals (in 2011 and 2012)?
> What would it mean for an activity to generate that?
> 
> It seems that this approach is restoring the attributes as characterising, which
> we voted against.
> 
> 
> 
> What maybe you mean is that the model does not specify exactly which thing
> an entity characterises ( in addition, the characterisation is not expected to be
> complete), but in theory there is one and only one thing.
> 
> 
> Professor Luc Moreau
> Electronics and Computer Science
> University of Southampton
> Southampton SO17 1BJ
> United Kingdom
> 
> On 18 Jan 2012, at 14:02, "Myers, Jim" <MYERSJ4@rpi.edu> wrote:
> 
> > Darn English -
> > Each entity represents one full characterization of some thing (and
> > only one *Characterization* - they are unambiguous w.r.t. everything
> > being asserted) -
> >
> > While an entity is asserted to represent a characterization of one thing, it
> may characterize many things - paolo-in-the-chair characterizes paolo with his
> location fixed and it characterizes person-in-chair with the person fixed.
> > Until processes occur that destroy paolo-in-the-chair, paolo-in-the-chair
> specializes both and they are alternates.
> >
> > I'll try to respond about time again, but in short - I think we can/should
> frame the definitions in terms of events rather than time per se, and that this
> is necessary to make alternate work for use cases like Stian's. Sitting and
> standing events create/destroy paolo-in-the-chair... I have to think more, but
> one could potentially make alternate fully transitive by requiring us to create
> 'person-in-chair-with-name-paolo' as a specialization of person-in-chair and
> make this the alternateOf paolo-in-chair - they have the same lifetime and
> hence 'time'/differences in lifetime go away. Saying paolo and person-in-chair
> are alternates over some interval between the sitting and standing events
> seems like an alternate way to model that would cover the same ground -
> we'd have 'pseudotransitivity' (which I think can be defined clearly over
> intervals) but would need fewer specialized entities in the description...
> >
> > Jim
> >
> >
> >> -----Original Message-----
> >> From: Paolo Missier [mailto:Paolo.Missier@ncl.ac.uk]
> >> Sent: Wednesday, January 18, 2012 8:40 AM
> >> To: Luc Moreau
> >> Cc: Paolo Missier; Stian Soiland-Reyes; public-prov-wg@w3.org
> >> Subject: Re: complementOf -> viewOf: proposed text
> >>
> >> Luc
> >>
> >> On 1/17/12 6:59 PM, Luc Moreau wrote:
> >>> Hi Paolo, Stian
> >>>
> >>> To answer the transitivity question, we need to answer a question.
> >>> Can an
> >> entity characterise different things? If yes, I agree transitivity
> >> does not necessarily hold. If no, transitivity holds.
> >> from Jim's earlier comprehensive message, I gather that the answer is not:
> >>> each entity therefore  represents one full characterization  of some
> >>> thing
> >> (and only one - they are unambiguous w.r.t. everything being asserted) .
> >> but also please see my reply to Jim, with the suggestion that
> >> temporal overlaps ought to be reintroduced. But we've been there
> >> before: entity records do not mention temporal validity. So if
> >> temporal overlaps are an argument against transitivity in this case,
> >> then I believe the best we can do is remove transitivity and do a bit
> >> of hand-waiving when asked about it :-)
> >>
> >> -Paolo
> >
Received on Thursday, 19 January 2012 13:31:13 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:58:11 UTC