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

Re: PROV-ISSUE-447: subactivity relation [prov-dm]

From: Timothy Lebo <lebot@rpi.edu>
Date: Tue, 4 Sep 2012 13:06:25 -0400
Cc: Provenance Working Group <public-prov-wg@w3.org>
Message-Id: <15A95312-B1AE-40D9-BEEF-30833FA9F385@rpi.edu>
To: "Miles, Simon" <simon.miles@kcl.ac.uk>
Simon,

On Sep 4, 2012, at 12:55 PM, Miles, Simon wrote:

> Hi Luc,
> 
> OK, thanks for clarifying. Though I still don't find it clear to both say that it is out of scope because there is inadequate state of the art and that dct:hasPart, which is part of the state of the art, can be used. Don't we need to choose one of the two?
> 
> I don't think we have to equate adding a term to PROV with defining our own new term. We can recommend using an existing term in a way that only makes sense in considering PROV data (e.g. dct:hasPartOf relating super- to sub-activities). I don't think we also have to justify it being out of scope, as then why would we make that recommendation?


What we have to justify is that any possible inferences of using some:isPartOf are out of scope.


> 
> People might wonder why we recommend using dct:hasPartOf for activities but to use prov:hadMember relation for entities, when dct:hasPart is defined in terms of relating "resources". I don't have an answer to this, but maybe it's not an issue.


Good point. I think we should be rather agnostic about _which_ part Of property that could be used instead of choosing one.
in a way, prov:hadMember exists b/c we didn't choose an existing one.

For this, we could name the property, or just state "there are properties out there that can be used for part hood".

Can we use prov:hadMember to bin a sub activity into a super activity ;-)


> 
> It's an interesting point about relating subactivities to subplans, though I didn't take it to be part of the original comment.

I didn't either.

> I think we already make clear that PROV does not touch on the structure of plans, which is technology-specific data content and not provenance, so it should be clearly out of scope.


Agreed. But there are still "possible PROV inferences among sub activities and prov:X" that one would reasonably want to make, and we didn't provide that solution.


-Tim

> 
> thanks,
> Simon
> 
> Dr Simon Miles
> Senior Lecturer, Department of Informatics
> Kings College London, WC2R 2LS, UK
> +44 (0)20 7848 1166
> 
> Provenance: The bridge between experiments and data:
> http://eprints.dcs.kcl.ac.uk/1372/
> ________________________________________
> From: Luc Moreau [l.moreau@ecs.soton.ac.uk]
> Sent: 04 September 2012 17:21
> To: public-prov-wg@w3.org
> Subject: Re: PROV-ISSUE-447: subactivity relation [prov-dm]
> 
> Hi Simon,
> 
> On 04/09/12 17:06, Miles, Simon wrote:
>> Hello Luc, Tim,
>> 
>> OK. I'm not clear from your replies whether you are disagreeing with me or not. Luc's options were (1) add the relation or (2) justify why it's out of scope. Regardless of whether we know in what way the subactivity's provenance is part of the superactivity's provenance, the fact is that it is a part. If you are disagreeing with me, and saying we should go with (2), then what is the justification for it being out of scope?
> 
> I suggest (2).
> 
> wasSubactivityOf is out of scope for this standardization activity because
> there is not enough experience and previous state-of-the art on this
> subject.
> 
> 
> It is suggested that relations such as dcterms:hasPartof can be used for
> this purpose.
> 
> 
>> 
>> I read the content of Tim's mail as agreeing with me on option (1) and further suggesting that the relation should be dc:hasPart, which seems plausible enough if the definition fits. Luc agreed with this. But both your emails sounded like you were disagreeing with me and so want option (2), so I'm quite confused... Which option are you proposing?
> 
> I read Tim's email as saying that dcterms:hasPartOf can be used for
> that. It's a dcterms property, not a prov property.
> 
>> 
>> I don't think the references Luc provided are relevant, as, unless I'm misreading, they are about decomposition of tasks in plans, not about how two activities related in their histories.
> Exactly, so we also need to relate subactivities to subplans potentially ...
> 
> Luc
> 
>> thanks,
>> Simon
>> 
>> Dr Simon Miles
>> Senior Lecturer, Department of Informatics
>> Kings College London, WC2R 2LS, UK
>> +44 (0)20 7848 1166
>> 
>> Provenance: The bridge between experiments and data:
>> http://eprints.dcs.kcl.ac.uk/1372/
>> ________________________________________
>> From: Luc Moreau [l.moreau@ecs.soton.ac.uk]
>> Sent: 04 September 2012 16:18
>> To: public-prov-wg@w3.org
>> Subject: Re: PROV-ISSUE-447: subactivity relation [prov-dm]
>> 
>> Hi Tim,
>> 
>> On 04/09/12 15:58, Timothy Lebo wrote:
>>> On Sep 4, 2012, at 10:45 AM, Miles, Simon wrote:
>>> 
>>>> Hello Luc,
>>>> 
>>>> I'd support option 1. I think it has relevance to provenance, in that if you say A wasSubactivityOf B, then any information about the provenance of A is part of the provenance of B.
>>>> 
>>>> I can't recall why we said it was out of scope before. I think the term "wasSubtaskOf" may imply the wrong thing, i.e. that the statement is about what was planned rather than (or as well as) what occurred, and other vocabularies must already cover this non-provenance assertion.
>>>> 
>>> I don't think that the relation from sub activity to activity was the concern for scope -- it was the consequences of what "provenance of A is part of the provenance of B" would mean.
>> Agreed.
>> 
>>> I would be content with a "silent" relation that does nothing but tuck an activity as part of another (why not just use dcterms:hasPart)?
>> Yes, good suggestion. This means that this relation is not part of PROV.
>> Our FAQ/response could
>> make this clear, and we could suggest the use of dcterms:hasPart.
>> 
>>> I don't have the ability or energy to revisit all of the discussions that we've had to see how any assertion should or should not  apply to an activities' sub activities.
>> The change is not trivial, and should have been considered at the
>> beginning of the design. It was not because it was not in our charter.
>> 
>> Luc
>> 
>>> -Tim
>>> 
>>> 
>>> 
>>> 
>>>> thanks,
>>>> Simon
>>>> 
>>>> Dr Simon Miles
>>>> Senior Lecturer, Department of Informatics
>>>> Kings College London, WC2R 2LS, UK
>>>> +44 (0)20 7848 1166
>>>> 
>>>> Provenance: The bridge between experiments and data:
>>>> http://eprints.dcs.kcl.ac.uk/1372/
>>>> ________________________________________
>>>> From: Luc Moreau [l.moreau@ecs.soton.ac.uk]
>>>> Sent: 04 September 2012 14:57
>>>> To: public-prov-wg@w3.org
>>>> Subject: Re: PROV-ISSUE-447: subactivity relation [prov-dm]
>>>> 
>>>> Dear all,
>>>> 
>>>> I would like to kickstart discussion on this public comment.
>>>> This has already been asked on several occasions, and this has previously
>>>> been raised on the mailing list.
>>>> 
>>>> I essentially see two options:
>>>> 1. We change the model and add a sub-activity relation.
>>>> 2. We don't change the model, but we come with a good justification for not
>>>>      changing it.  In particular, we previously said this was out of
>>>> scope. Perhaps,
>>>>      we could point to some vocabularies already doing this.
>>>> 
>>>> What are your views?
>>>> Regards,
>>>> Luc
>>>> 
>>>> On 06/07/12 18:12, Provenance Working Group Issue Tracker wrote:
>>>>> PROV-ISSUE-447: subactivity relation [prov-dm]
>>>>> 
>>>>> http://www.w3.org/2011/prov/track/issues/447
>>>>> 
>>>>> Raised by: Timothy Lebo
>>>>> On product: prov-dm
>>>>> 
>>>>> There is a thread discussing the issue raised by Sutra at http://www.w3.org/mid/CAJCyKRqtC47OWc_rDRhFcQGdJ-yy2toQBCguUywFGZpHO5Q8Jw@mail.gmail.com
>>>>> 
>>>>> The original email:
>>>>> 
>>>>> On Fri, Jul 6, 2012 at 2:45 PM, Satrajit Ghosh <satra@mit.edu> wrote:
>>>>> hello,
>>>>> 
>>>>> i was discussing this with luc and based on his feedback thought it might be
>>>>> useful to bring this up on the list.
>>>>> 
>>>>> ----
>>>>> question:
>>>>> how do you encode that a certain activity "emailing a letter" happened
>>>>> during another activity "a meeting"?
>>>>> 
>>>>> for example we conduct research studies/projects.
>>>>> 
>>>>> activity(p1, [prov:type='ex:Project'])
>>>>> activity(p2, [prov:type='ex:MRIScanning', ex:session=1])
>>>>> activity(p3, [prov:type='ex:MRIScanning', ex:session=2])
>>>>> 
>>>>> how would i encode that this activity p2 and p3 were conducted during p1?
>>>>> how would i encode p3 followed p2?
>>>>> 
>>>>> 
>>>>> luc's response:
>>>>> Regarding your question, there may be a few options:
>>>>> you could add time information to your activities. This will help you
>>>>> understand their ordering.
>>>>> 
>>>>> Alternatively, if you want an explicit dependency in your graph, then p2 may
>>>>> generate something
>>>>> that starts p3, and/or is consumed by p3
>>>>> 
>>>>> Finally, prov doesn't have relations between activities, to express their
>>>>> nesting, etc. It's important
>>>>> but we felt this is not specific to provenance, but to process executions.
>>>>> ----
>>>>> 
>>>>> it's the last point on this response that i was not completely sure about.
>>>>> why "relations between activities" is "not specific to provenance, but to
>>>>> process executions."
>>>>> 
>>>>> in the above example, one could say:
>>>>> 
>>>>> wasSubtaskOf(p2, p1)
>>>>> wasSubtaskOf(p3, p1)
>>>>> wasFollowedBy(p2, p3)
>>>>> 
>>>>> any clarification as to why such relations would be outside the realm of
>>>>> provenance would be much appreciated.
>>>>> 
>>>>> cheers,
>>>>> 
>>>>> satra
>>>>> 
>>>>> 
>>>>> 
>>>> --
>>>> 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
>>>> 
>> --
>> 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
> 
> --
> 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
> 
Received on Tuesday, 4 September 2012 17:06:58 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 4 September 2012 17:06:58 GMT