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

Re: PROV-ISSUE-4 (define-agent): Definition for Concept 'Agent' [Provenance Terminology]

From: Graham Klyne <GK@ninebynine.org>
Date: Wed, 08 Jun 2011 09:06:07 +0100
Message-ID: <4DEF2D6F.7020709@ninebynine.org>
To: Simon Miles <simon.miles@kcl.ac.uk>
CC: Provenance Working Group WG <public-prov-wg@w3.org>
I agree with Simon, especially para 2. (Am I making a habit of this?)

#g
--

Simon Miles wrote:
> Hi Khalid, all,
> 
> These are really good points, and makes me think:
>  - If an agent starts or stops a process execution (as in Satya's
> definition) then why is that any different to one process execution
> initiating or terminating another process execution?
>  - Is the action of a person a process execution? If so, and the
> action is adequately described, then haven't we just specified the
> involvement and identity of the person in the provenance?
> 
> I agree with Jun that if we want to describe the provenance of an
> agent then we are treating it as a "resource", but Khalid's point that
> we still need to say they are the same (this agent = this resource)
> seems correct. On a separate thread, Jim said (if I read correctly)
> that the idea of agent may not be needed if mutable resources were
> allowed, as a person is just a mutable resource. We can take an
> invariant view of a person in provenance, i.e. assume they don't
> change in any way we consider significant.
> 
> The significant differences of an agent from an execution or resource
> seem to be that we may want to express its involvement in an execution
> without saying exactly by what actions it was involved, and that it
> seems odd to say that a process execution "used" the person who was
> involved with it as we would with resources (which we could also say
> are "involved" with the execution).
> 
> Thanks,
> Simon
> 
> On 7 June 2011 19:17, Khalid Belhajjame <Khalid.Belhajjame@cs.man.ac.uk> wrote:
>> Hi Daniel, Jun and Satya,
>>
>> I think this is a good point. I was wondering before this thread start,
>> whether the same "thing" can be treated as "agent" and "process execution".
>> For example a process execution pe_1 may initiates the execution of
>> another process execution pe_2, in which case, pe_1 can be viewed as an
>> agent.
>>
>> So a legitimate question that we may need to create an independent issue
>> for is: "is it possible for the same 'thing' to be 'invariant' and
>> 'agent' and 'process execution'?"
>>
>> My first thought is that the notion of "view or account" will be useful
>> in this respect.
>>
>> Thanks, khalid
>>
>> On 07/06/2011 17:35, Jun Zhao wrote:
>>> Satya,
>>>
>>> Would you also expect agents to have states, aka resources?
>>>
>>> I would prefer to keep agent simple. That's why I suggest to treat
>>> agents as resources, if you want to talk about provenance of that sort
>>> of things. In this way, we can reuse of a lot of structure associated
>>> with resources.
>>>
>>> This is just one way of modelling. Of course, if we do agree to have
>>> states for agents, we will need a different way to represent it in the
>>> model.
>>>
>>> I am open to discussions.
>>>
>>> -- Jun
>>>
>>> On 07/06/11 17:17, Satya Sahoo wrote:
>>>> Hi Daniel and Jun,
>>>>> After reading your definitions I was wondering if an "agent" or a
>>>> "controller" could
>>>> also have provenance.
>>>>
>>>> Agents can also have provenance  - in sensor networks the provenance of
>>>> the sensor itself will describe its manufacturer, date of manufacture
>>>> etc.
>>>>
>>>>> if the newspaper or a service is an agent, then it would be helpful
>>>> to know who created it, when, what tools were used for doing so,etc.
>>>>
>>>> If newspaper is treated as an agent in the sense of a corporate entity,
>>>> e.g. NYT - we can describe its actions - NYT sued NYC, NYT publishes
>>>> four dailies etc. and its provenance - NYT was founded/created by Henry
>>>> Raymond in 1851 etc.
>>>>
>>>> If you meant newspaper as an edition/publication of NYT (today's
>>>> edition), then it would be a "resource". I am not sure what you meant by
>>>> service?
>>>>
>>>> Thanks.
>>>>
>>>> Best,
>>>> Satya
>>>>
>>>> On Tue, Jun 7, 2011 at 11:57 AM, Jun Zhao <jun.zhao@zoo.ox.ac.uk
>>>> <mailto:jun.zhao@zoo.ox.ac.uk>> wrote:
>>>>
>>>>     Hi Daniel,
>>>>
>>>>     I would expect to have descriptions about agents.
>>>>
>>>>     When providing provenance descriptions about newspaper or services,
>>>>     it might be more appropriate to treat them as "resources".
>>>>
>>>>     -- Jun
>>>>
>>>>
>>>>
>>>>     On 07/06/11 16:19, Daniel Garijo wrote:
>>>>
>>>>         Hi Satya, Khalid, all
>>>>         After reading your definitions I was wondering if an "agent"
>>>> or a
>>>>         "controller" could
>>>>         also have provenance. For example, if the newspaper or a service
>>>>         is an
>>>>         agent, then it would be
>>>>         helpful to know who created it, when, what tools were used for
>>>>         doing so,
>>>>         etc.
>>>>
>>>>         If they could have provenance, then they would also become some
>>>>         sort of
>>>>         "resources"
>>>>         too, wouldn't they?
>>>>
>>>>         What do you think?
>>>>
>>>>         Best,
>>>>         Daniel
>>>>
>>>>         2011/6/6 Khalid Belhajjame <Khalid.Belhajjame@cs.man.ac.uk
>>>> <mailto:Khalid.Belhajjame@cs.man.ac.uk>
>>>> <mailto:Khalid.Belhajjame@cs.man.ac.uk
>>>> <mailto:Khalid.Belhajjame@cs.man.ac.uk>>>
>>>>
>>>>
>>>>
>>>>             Hi,
>>>>
>>>>             I added a definition of the concept "agent".
>>>>
>>>>             Definition: An agent is a (physical or digital) entity that
>>>>         controls
>>>>             one or multiple process executions
>>>>
>>>>             - The newspaper, the blogger and the government portal are
>>>>         examples
>>>>             of agents
>>>>
>>>>             I would prefer to use the term “controller” instead of
>>>>         “agent”. To
>>>>             me, the term “agent” is a bit vague.
>>>>
>>>>             Thanks, khalid
>>>>
>>>>
>>>>             On 20/05/2011 08:04, Provenance Working Group Issue Tracker
>>>>         wrote:
>>>>
>>>>                 PROV-ISSUE-4 (define-agent): Definition for Concept
>>>> 'Agent'
>>>>                 [Provenance Terminology]
>>>>
>>>>         http://www.w3.org/2011/prov/track/issues/4
>>>>
>>>>                 Raised by: Luc Moreau
>>>>                 On product: Provenance Terminology
>>>>
>>>>                 The Provenance WG charter identifies the concept 'Agent'
>>>>         as a
>>>>                 core concept of the provenance interchange language
>>>> to be
>>>>                 standardized (see
>>>>         http://www.w3.org/2011/01/prov-wg-charter).
>>>>
>>>>                 What term do we adopt for the concept 'Agent'?
>>>>                 How do we define the concept 'Agent'?
>>>>                 Where does concept 'Agent' appear in ProvenanceExample?
>>>>                 Which provenance query requires the concept 'Agent'?
>>>>
>>>>                 Wiki page: http://www.w3.org/2011/prov/wiki/ConceptAgent
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>>
>>>
>>>
>>
>>
>> ______________________________________________________________________
>> This email has been scanned by the MessageLabs Email Security System.
>> For more information please visit http://www.messagelabs.com/email
>> ______________________________________________________________________
>>
> 
> 
> 
Received on Wednesday, 8 June 2011 11:12:42 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 13:06:31 GMT