W3C home > Mailing lists > Public > public-semweb-lifesci@w3.org > June 2007

RE: [Obo-relations] Advancing translational research with the Semantic Web (Not clear about definition of <is_location_of_process>)

From: Smith, Barry <phismith@buffalo.edu>
Date: Fri, 01 Jun 2007 14:15:37 -0400
To: "Kashyap, Vipul" <VKASHYAP1@PARTNERS.ORG>
Cc: "William Bug" <William.Bug@DrexelMed.edu>, "Smith, Barry" <phismith@buffalo.edu>,Cc:
Message-Id: <20070601182249.DDHG16681.mta13.adelphia.net@LENOVO-B1AD6243.buffalo.edu>

At 01:36 PM 6/1/2007, you wrote:

> > The life process, for instance, in cryogenics?
>
>[VK] Would you consider the life process to be suspended when a person is in a
>coma? And reactivated when he emerges from a coma?

No. (Because this is what biology tells us.)

> > >- An interesting corollary is that the execution of a program needs to be
> > >distinguished from a program (please feel free to fill in the biological
> > >equivalents).
> >
> > This is the basis of BFO's discussion between realizable entities
> > such as functions and the processes which are their realizations; the
> > former are continuants, the latter are occurrents.
> >
> > >- A process comes into existence only when a computer program executes.
> >
> > A process of a certain kind ...
>
>Let me try to use some of your terminology (please correct me if there are any
>errors):
>- A process is an occurrent
>- It is a realization (or manifestation?) of a function or a program

only certain very special kinds of processes

>- Because it is an occurrent "it has temporal parts and unfolds itself in
>multiple phases"

Good


>So process phase = process state...

No. If we take a snapshot of the bearer of the process at a time (the 
device, in this case), then we get what we might reasonably call the 
state of the device at that time. But process phases last intervals 
of time, and during any such interval there might be an arbitrary 
number of such states.

>Return to the program-process: It can be in an activation phase, 
>running phase,
>suspended phase, terminated phase (vhere it ceases existence)...

This sounds better for running phase; the activation phase would not 
be a phase (since it presumably is instantaneous); rather it is the 
boundary of a phase (The start of a race is not a phase of the race; 
it is the boundary of the race.)

It doesn't work for 'suspended phase' or 'terminated phase' at all, I 
think. (And I hope it is clear why, at least for the latter.)

If you mean by 'termination' the dual of activation, namely the 
instantaneous boundary of a process when it cases, then this, too, is 
not a phase, since it has no temporal duration.

>At the very least, since an occurrent either exists/doesn't exist at a given
>point in time, a process at least has two states:
>- Existence
>- Non-existence

NO NO NO.
There are (by definition of 'non-existence') no non-existent entities.
Hence there are no non-existent processes.
Hence there are no processes which have the putative state of non-existence.

>Different kinds of processes like life processes and computer processes could
>have more specialized states?

Yes of course.

> > >  For instance
> > >the process1 = execution of the web browser program; waits-for
> > >messages from process2 = execution of the web server program.
> >
> > Again, you are confusing the device which executes with the process
> > which is the execution. The device waits.
>
>[VK] Let's consider the following scenario:
>A process can be swapped out of memory and another process can be swapped into
>memory for execution.

I think the problem truly is that you are using the word 'process' in 
a way quite different from the way it is standardly used in ontology 
circles. What you mean is something like an algorithm. Algorithms 
have executions, and it is these executions which I would prefer to 
call processes. Let's use 'VKprocess' to mean: that which gets 
executed when a computer program like the one you just described is 
run. Then this

A VKprocess can be swapped out of memory and another VKprocess can be 
swapped into
memory for execution.

makes sense, and a VKprocess is a continuant, not an occurrent.

>In that case the process is still waiting, while the
>device (because it's executing another process) is not.

The VKprocess is waiting. BSprocesses do not wait.

>I would propose that in the above scenario, it's the process which is waiting
>and not the device?

The VKprocess, yes.

>Also, in general one view of a process could be:
>Process = Participants (Actors) + Actions (Tasks) + Temporal Parts (Phases?)

what is the meaning of '+' here?

>Actions could lead to phase transition, e.g., migration of a process from one
>device to another.

BSprocesses do not migrate.

>There are two ways then of defining <process-location>
>1. The location of (All) participants of the process at a particular point in
>time
>2. The location of a Task of the process at a particular point in time.
>
>Obviously you have chosen (1), though (2) is an alternate way of defining
><process-location> and probably in the context of a given definition of a
>process they are equivalent. Note that I have removed references to 
>state/phase
>from (2).
>
>Why would you believe (if you do), that (1) is a more well-founded definition
>than (2)?

Can you rephrase this last bit using the 'VKprocess' terminology?
BS 
Received on Friday, 1 June 2007 18:23:00 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 18:00:48 GMT