Re: Fwd: provenance working group draft specs - ready for review

Looking at the proposal for Activity made in the context of, 
it seems that the mappings, as mentioned by Dan Brickley, is not a 
simple 1 to 1 mapping. The notion of Activity in seems to 
encompasses the notion of action, which I think is inline with was we 
call Activity in prov, plus information about the performer (i.e., 
Agent), input items (i.e., used entities), and the results (i.e., 
generated entities). Despite the complexity of the mappings, I think 
they can be encoded in a straightforward manner using, e.g., SPARQL.

thanks, khalid

On 05/05/2012 12:39, Paul Groth wrote:
> Hi All,
> Below is an exchange with Dan Brickley about the opportunity to align
> with around activity.
> As I say below, the models align well. I think this is something that
> we as a group should pursue.
> cheers
> Paul
> ---------- Forwarded message ----------
> From: Dan Brickley<>
> Date: Fri, May 4, 2012 at 9:44 PM
> Subject: Re: provenance working group draft specs - ready for review
> To: "Groth, P.T."<>
> Cc: SemWeb meetings<>
> On 4 May 2012 21:41, Groth, P.T.<>  wrote:
>> Hi Dan,
>> Thanks for the quick feedback.
>> I think the Activity in Web schema aligns nicely with prov and I would
>> be keen on doing that. I think all the terms you listed have
>> correspondence in prov.With respect to roles (butcher, baker) this is
>> supported through prov:role that is attached to the edges one uses to
>> connect a thing or agent to an activity.
> Ah, I missed that; thanks!
>> Your right that prov:Entity acts as a place holder for things
>> connected to provenance. Although, we do assign some additional
>> semantics to prov:Entity when we add constraints to the model.
>> Would you mind sending your comments to the whole WG at
>> or have me forward them to the group?
> Feel free to fwd them directly,
> cheers,
> Dan
>> Thanks
>> Paul
>> On Fri, May 4, 2012 at 1:19 PM, Dan Brickley<>  wrote:
>>> On 4 May 2012 12:00, Groth, P.T.<>  wrote:
>>>> Hi All,
>>>> I know a number of you are working on provenance related things. The
>>>> Provenance Working Group just released a major set of working drafts.
>>>> This is cohesive set of drafts and has really settled down.
>>>> I would appreciate you looking into them and providing comments. Also,
>>>> if anyone plans to implement these specs I would appreciate hearing
>>>> about it.
>>>> To get into the specs, we've prepared some introductory blog posts.
>>>> Please have a look.
>>>> PROV: synchronized and ready for your input
>>>> -
>>>> The PROV ontology – an update
>>>> -
>>>> What is new in the Fourth Working Draft of the PROV provenance model?
>>>> -
>>>> Again, the group is looking for your feedback and is looking to
>>>> finalize the interchange model soon.
>>> Thanks for the overviews, the list of docs is a little daunting otherwise.
>>> Looking at the activities piece, around
>>> I wonder if
>>> there's scope for making sure new activities/action
>>> vocabulary covers these use cases, even if the terminology is not
>>> exactly 1:1? There's a proposal at
>>> ... key properties
>>> for an Activity there are
>>> performer -- who performed the action. Person or Organization.
>>> location -- where the action was performed.
>>> startDate, endDate, duration -- when the action was performed.
>>> action -- the action which was performed. URL or Action [new type
>>> defined below].
>>> item -- upon what the action was performed. Thing.
>>> result -- how the world has changed because of the action. Thing.
>>> Often Comment, Review, Photograph, BlogPost, ItemList, etc., but
>>> complex activity-specific metadata may also be represented with an
>>> activity-specific result item.
>>> I see a few more details are in the Prov model but not in this
>>> proposal, and there are event models like
>>> not far away too. How do
>>> you handle the roles that people play (butcher, baker,
>>> candlestickmaker, ...), indicators for verbs etc? From the example it
>>> seems the work is put onto the specific activity type, but then I
>>> don't see how to say that alice was the 'sound engineer', and bob was
>>> the 'lead compositor'? Is that kind of detail in scope? I wonder how
>>> much scope there is for a common model here.
>>> The only other thing that leaps out from a quick look is
>>> "prov:Entity". If someone tells me that some mystery object X is a
>>> prov:Entity, have I learned anything about X? Are there any things
>>> that aren't a prov:Entity? I get feeling it's a kind of place marker
>>> rather than a distinctive type, a way of hinting "this description is
>>> talking about the provenance of these things..."?
>>> cheers,
>>> Dan
>> --
>> --
>> Dr. Paul Groth (
>> Assistant Professor
>> Knowledge Representation&  Reasoning Group
>> Artificial Intelligence Section
>> Department of Computer Science
>> VU University Amsterdam

Received on Saturday, 5 May 2012 12:42:34 UTC