W3C home > Mailing lists > Public > public-prov-wg@w3.org > July 2011

RE: PROV-ISSUE-34: Section 4: definition of "Agent"

From: Myers, Jim <MYERSJ4@rpi.edu>
Date: Sun, 24 Jul 2011 10:58:05 -0400
Message-ID: <B7376F3FB29F7E42A510EB5026D99EF20546895B@troy-be-ex2.win.rpi.edu>
To: Khalid Belhajjame <Khalid.Belhajjame@cs.man.ac.uk>, Paul Groth <p.t.groth@vu.nl>
CC: <khalidb@cs.man.ac.uk>, <public-prov-wg@w3.org>
I think the choice being made is really on the part of the asserter.
There's nothing in the model that says I can't use a really long-lived
BOB for the agent, i.e. one defined as 'Alice while she's alive'.
There's also nothing to stop an asserter from creating a long-string of
BOBs representing Alice in 'sleepy' and 'rested' states that would allow
one to start understanding whether Alice makes better decisions when
rested.

And this is a good thing - both are valid use cases and the model we're
discussing/refining would a) allow both, and b) let you connect the two
(IVPof) views.

-- Jim

> -----Original Message-----
> From: public-prov-wg-request@w3.org [mailto:public-prov-wg-
> request@w3.org] On Behalf Of Khalid Belhajjame
> Sent: Sunday, July 24, 2011 6:43 AM
> To: Paul Groth
> Cc: khalidb@cs.man.ac.uk; public-prov-wg@w3.org
> Subject: Re: PROV-ISSUE-34: Section 4: definition of "Agent"
> 
> 
> Hi Paul,
> 
> On 23/07/2011 15:36, Paul Groth wrote:
> > Hi Khalid,
> >
> > I think the definition of agents as a BOB is fine. The definition
> > doesn't prevent you from having things that change about an agent
> > outside the things that you defined as fixed.
> 
> I agree. What I raised was not really an issue, it was more a
consequence of the
> choice we've made that I thought we should be aware of. Actually, the
more I
> think about it, the more I find that having multiple BOBs that are
associated
> with the same process execution, is fine, even when the different BOB
refers to
> the same human or system.
> This is because defining an agent as a BOB, as opposed to an entity,
give us
> more information (provenance) about the properties of the agent when
> controlling the process execution.
> 
> Thanks, khalid
> 
> >
> > So for alice:
> >
> > If the attributes that characterized Alice, were her first name,
last
> > name, and facebook url then changing her profession wouldn't matter.
> >
> > If the core attributes change about how we identify alice I think
it's
> > fair enough to require that those changes be reflected in
provenance.
> >
> > It's almost as if you were changing who you believed was controlling
> > the process and indeed if a core attribute of the person changes
this
> > is important information to know.
> >
> > Does that make sense?
> >
> > Paul
> >
> >> According to the definition in the Provenance Model initial draft
"An
> >> agent represents a characterized entity capable of activity".
> >>
> >> My interpretation of this definition is that "an agent is a BOB".
If
> >> that is the case, then one of the consequences is that we may need
to
> >> associate a given process execution with multiple Agents that refer
> >> to the same human (system). To illustrate this, consider a long
> >> running process execution that is controlled by Alice, and consider
> >> that one of the attribute characterizing Alice, e.g., grade,
changed,
> >> e.g., she was promoted, in that case, we will need to create a new
> >> BOB (that characterizes Alice) and associate it with the running
> >> process execution. Are we happy with this?
> >>
> >> Khalid
> >
> >
> 
Received on Sunday, 24 July 2011 14:58:49 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 13:06:37 GMT