W3C home > Mailing lists > Public > public-prov-wg@w3.org > February 2012

Re: PROV-ISSUE-237 (TLebo): Rename Relation to Involvement [prov-dm]

From: Timothy Lebo <lebot@rpi.edu>
Date: Mon, 6 Feb 2012 20:44:21 -0500
Cc: public-prov-wg@w3.org
Message-Id: <4ABF59F4-D4D5-4E2B-99E0-687AF4DBAA06@rpi.edu>
To: reza.bfar@oracle.com

On Feb 6, 2012, at 8:21 PM, Reza B'Far (Oracle) wrote:

> Tim -
> 
> I also saw your other note to Luc and Paolo.  I would suggest, based on your logic, that your proposal of "involve" is replaced with "participate"

The wg wrestled with "participate" around the first F2F. The problem with it is that it insinuates too much agency.
It was replaced with "wasAssociatedWith" which _is_ the one that bestows agency (but might not for much longer).



> and then somehow also add the decoration of whether it's active or passive participation.  

I very much like the bipartition of active and passive. That's the center of what we keep spinning around.

To me, passive participation is involvement.
The ball I threw at Khalid's head was involved in the Assault activity.
I find it hard to see the ball as a participant.


> 
> I think involve is used (at least colloquially) in vague ways... like "involved" can mean "complex" or "complicated" in some contexts, etc.


In some contexts, it can.

If  :thing rdf:type prov:Involved, I could see how that could be interpreted as "complex".

But as a binary (or n-ary) relation, I find it hard to interpret 

:email_writing prov:involved :the_e_key  .

as "complicated"

-Tim

What I'm shooting for in PROV-O:

prov:involved rdfs:domain [ owl:unionOf ( prov:Activity prov:Entity ) ]



> 
> Best.
> 
> On 2/6/12 5:09 PM, Timothy Lebo wrote:
>> 
>> Hi, Paolo,
>> 
>> On Feb 6, 2012, at 5:55 PM, Paolo Missier wrote:
>> 
>>> Tim,
>>>  I am not sure I understand. The term "relation" is entirely standard in data modelling,
>> I would say because we are not creating a metamodeling language like UML or ERD. We're only making a model.
>> So we shouldn't be using the general term for what we're doing.
>> 
>>> as well as in set theory. "association" is used instead in UML and I wouldn't object to that. But why do we need to spend time looking for alternatives?
>> Acknowledged. Time is short.
>> However, time spent making this model easier to understand is worthwhile.
>> 
>> PROV is offering a very limited set of relations, and I find the disparity in breadth to be dissonant.
>> In talking about the model with others, I have found that they agree.
>> 
>> -Tim
>> 
>> 
>> 
>>> --Paolo
>>> 
>>> 
>>> 
>>> On 2/6/12 9:32 PM, Luc Moreau wrote:
>>>> Hi Tim,
>>>> 
>>>> I am keen to replace 'relation' (and 'element') by more appropriate names.
>>>> 
>>>> I am not sure why 'involvement'?  involvement in what?
>>>> 
>>>> How appropriate is it for alternateOf?
>>>> 
>>>> Thanks,
>>>> Luc
>>>> 
>>>> On 06/02/12 21:01, Provenance Working Group Issue Tracker wrote:
>>>>> PROV-ISSUE-237 (TLebo): Rename Relation to Involvement [prov-dm]
>>>>> 
>>>>> http://www.w3.org/2011/prov/track/issues/237
>>>>> 
>>>>> Raised by: Timothy Lebo
>>>>> On product: prov-dm
>>>>> 
>>>>> I propose to rename "Relation" in PROV-DM to "Involvement" because "Relation" is too broad and a provenance interchange should limit itself to how agents, activities, and entities were involved with one another as the lead to some result.
>>>>> 
>>>>> Relations other than involvements should be out of scope for provenance interchange (and seem to be already be handled with the attribute-values).
>>>>> 
>>>>> Thanks,
>>>>> Tim
>>>>> 
>>>>> 
>>>>> 
>>>>> 
>>> 
>>> -- 
>>> -----------  ~oo~  --------------
>>> Paolo Missier - Paolo.Missier@newcastle.ac.uk, pmissier@acm.org
>>> School of Computing Science, Newcastle University,  UK
>>> http://www.cs.ncl.ac.uk/people/Paolo.Missier
>>> 
>>> 
>>> 
>> 
Received on Tuesday, 7 February 2012 01:47:27 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 13:06:56 GMT