Re: PROV-ISSUE-66 (is-execution-a-bob): Why is process execution not defined as a characterised entity? [Conceptual Model]

Hi Jim,
> I would not consider use and generation to fit the term participation –
they are 1-way terms whereas participation implies (to me) some
back-and-forth interaction
I was interpreting participation as a 1-way term (irreflexive?), since an
agent may participate in a process, but the process does not participate in
agent?

Maybe for the interaction property, we should define "interactsWith"
property?

>  something vague like affectedBy seems like the parent idea.
We can interpret this to be 1-way also, for example A isAffectedBy B, but B
is not affectedBy A?

>  I do see participates being consistent both with things that are the
‘substrate’ of a PE and agents/control – both have that interaction flavor.
But again – the key thing is to have something distinct from use and
generation in the model. Having a superclass for all of them sounds fine if
we see the need/use case – probably a useful convenience anyway for
querying.
I agree, should we consider interactsWith property as a potential candidate?

Thanks.

Best,
Satya



On Tue, Aug 2, 2011 at 6:01 PM, Myers, Jim <MYERSJ4@rpi.edu> wrote:

> >> What would you call the type of participation that is neither used nor
> generation? (I suggested reserving the term participation for this subset,
> but as long as we can distinguish it in the model, the name is less
> important.)****
>
> >We had earlier discussed "involvedIn" - I am not suggesting that we use
> the term but it is a potential candidate. Since use and generation also are
> "kinds of participation", it may be more natural to consider it as a parent
> property?****
>
> ** **
>
> ** **
>
> I would not consider use and generation to fit the term participation –
> they are 1-way terms whereas participation implies (to me) some
> back-and-forth interaction – something vague like affectedBy seems like the
> parent idea. I do see participates being consistent both with things that
> are the ‘substrate’ of a PE and agents/control – both have that interaction
> flavor. But again – the key thing is to have something distinct from use and
> generation in the model. Having a superclass for all of them sounds fine if
> we see the need/use case – probably a useful convenience anyway for
> querying.****
>
> ** **
>
> Cheers,****
>
> Jim****
>