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

Re: PROV-ISSUE-91 (what-to-define-for-location): what should we define under the heading 'location' [Conceptual Model]

From: Stian Soiland-Reyes <soiland-reyes@cs.manchester.ac.uk>
Date: Thu, 1 Dec 2011 09:23:03 +0000
Message-ID: <CAPRnXtmaJr7k6icys-1Gc6s8WhPoZFZE=OsUhQZBt_6WH5x5Rg@mail.gmail.com>
To: Paul Groth <p.t.groth@vu.nl>
Cc: Luc Moreau <L.Moreau@ecs.soton.ac.uk>, "public-prov-wg@w3.org" <public-prov-wg@w3.org>
I agree - what is important from a provenance perspective would be the
property linking something WITH a location. How that location is
specified is beyond our knowledge - and we should allow any kind of
'location', not just "geo-spatial". (The filename examples contradicts
this already).


I see this along the same lines as the anchor for Recipe/Plan.

It comes to the prime questions of provenance.. WHAT happened, WHEN
did it happen, WHERE did it happen and WHO did it.

We've got WHAT and WHO pretty well covered - WHEN is tentatively
covered (PROV-O references time.owl which allows most WHEN
expressions), but WHERE would be lost if we remove Location.

I think both entities and activities MAY have a location specified. It
is of course not a requirement that the location is fixed with regards
to the Earth's surface, for instance a card transaction could take
place on a moving train between London and Paris, or a virtual
purchase takes place "on the website", and a file is located on a
certain filesystem path *and* a certain disk sector *and* on a certain
USB disk in someones pocket.

I don't think it should either be a requirement that locations are
consolable, e.g. if activity A uses entity e1 and generates e2 - all
three of these could have "different" locations. So I don't think
there will be any particular reasoning you can do about the location -
it is still very much a useful concept that would be common to many
provenance scenarios, just like agents and time.


In the current PROV-O FPWD provo:hadLocation can be used either with
an provo:Entity or with an provo:QualifiedInvolvement (specifying
location of the use/generation/control)  - but not with the actual
ProcessExecution.


On Wed, Nov 30, 2011 at 17:39, Paul Groth <p.t.groth@vu.nl> wrote:
> Hi Luc
> I believe it is essential to have hooks in the model to well known an important provenance concepts. This includes location.
>
> Paul
>
>
> On Nov 30, 2011, at 12:59, Luc Moreau <L.Moreau@ecs.soton.ac.uk> wrote:
>
>>
>> Dear all,
>>
>> Following this September message, I am proposing to close this issue
>> formally.
>>
>> This said, location is so weakly specified in the model, that it begs
>> the question "what
>> does it do in the model at all". No specific data type, no specific
>> property, no specific
>> constraint, no specific inference.  As part of the clean up, we could
>> simply drop this
>> concept, and it wouldn't affect inter-operability!
>>
>> If you have views on this, please raise them as new issues.
>>
>> Regards,
>> Luc
>>
>> On 09/23/2011 11:50 AM, Luc Moreau wrote:
>>>
>>> We are proposing to close this issue, pending review.
>>> The latest version of the document states that:
>>>
>>> Location is an OPTIONAL attribute of entity expressions and process
>>> execution expressions.
>>>
>>> Feel free to reopen it, if you feel this is not right.
>>> Regards,
>>> Luc
>>>
>>> On 06/09/2011 12:47, Provenance Working Group Issue Tracker wrote:
>>>> PROV-ISSUE-91 (what-to-define-for-location): what should we define
>>>> under the heading 'location' [Conceptual Model]
>>>>
>>>> http://www.w3.org/2011/prov/track/issues/91
>>>>
>>>> Raised by: Luc Moreau
>>>> On product: Conceptual Model
>>>>
>>>> We have received little input and there has been barely any
>>>> discussion on location.
>>>> The current text is essentially copied from the wiki and is not
>>>> aligned with the rest of the model.
>>>>
>>>> So, can entities have location?  If so, should location appear as an
>>>> attribute of an entity?
>>>> So, should PIDM define some core attributes? or should this be left
>>>> to a generic profile?
>>>>
>>>> What about process executions? PIDM does not have "attribute" for
>>>> PEs. So, do we need to define a relation hasLocation?
>>>>
>>>> Is location unique for an entity/pe?
>>>>
>>>> Can people who have interest in location for provenance provide us
>>>> with some guidance, so that we can write something sensible for the
>>>> FPWD.
>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>
>> --
>> Professor Luc Moreau
>> Electronics and Computer Science   tel:   +44 23 8059 4487
>> University of Southampton          fax:   +44 23 8059 2865
>> Southampton SO17 1BJ               email: l.moreau@ecs.soton.ac.uk
>> United Kingdom                     http://www.ecs.soton.ac.uk/~lavm
>>
>>
>



-- 
Stian Soiland-Reyes, myGrid team
School of Computer Science
The University of Manchester
Received on Thursday, 1 December 2011 09:23:57 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:51:04 UTC