Re: PROV-ISSUE-184: Section 2.1.2 (PROV-DM version as on Nov 28th) [prov-dm]

Hi Satya,

On 12/07/2011 01:43 AM, Provenance Working Group Issue Tracker wrote:
> PROV-ISSUE-184: Section 2.1.2 (PROV-DM version as on Nov 28th) [prov-dm]
>
> http://www.w3.org/2011/prov/track/issues/184
>
> Raised by: Satya Sahoo
> On product: prov-dm
>
> Hi,
> The following are my comments regarding Section 2.1.2 of the PROV-DM document as on Nov 28, 2011.
>
> Section 2.1.2:
> 1. "Hence, in our conceptualization of the world, an instantaneous event, or event for short, happens in the world and marks a change in the world, in its activities and in its entities."
>
> Comment: According to this definition, an activity is also an event, since "it happens in the world and marks a change in the world".
> Section 5.3.1.1 states "A generation event may be, for example, the creation of a file by a program, the creation of a linked data set, the production of a new version of a document, and the sending of a value on a communication channel." So, generation event is being used as a synonym of an activity?
>
>    

I think you are now playing with words. Can you make suggestions on how 
to address this problem, since to start with
activity is not defined very precisely (and intentionally so!).

Later in a document, event is defined as boundary of an interval. I 
don't think we can mix the concept of boundary and interval.
We just need to get the right wording.

Are you saying that generation is an activity? But then how do you 
relate activities and entities they generate?


> In addition, Section 2.1.1 describes a world as "real or not" and includes "digital, conceptual etc." - an event in many scenario will not mark a change in "that" world. For example, person looks at the clock and continues in previous state - it is an event but what changed in the world?
>    

What changed in the world? the state of knowledge of the person who 
looked at the clock.

> 2. "events represent communications or interactions; they are assumed to be atomic and instantaneous."
>
> Comment: This is contradictory to what is stated in Section 5.3.1.2, "A usage event may be the consumption of a parameter by a procedure, the reading of a value on a port by a service, the reading of a configuration file by a program, or the adding of an ingredient, such as eggs, in a baking activity." The reading of a configuration file by program is not instantaneous? Adding of eggs in baking activity is not instantaneous? As I have commented on the previous version of DM, this is an artificial requirement for events to be instantaneous.
>    

It's not artificial: it's the consequence of being an interval 
boundary.  It's also common practice in process algebra.


> In addition, why is it required that events are atomic? Adding ingredient in a baking activity is never atomic - egg shells are broken, egg is stirred, egg mix is slowly added to the cake batter, while the batter is mixed continuously etc. Similarly, for events such as reading of file - access is checked, read lock is acquired, i-node is updated, reader process reads content of file etc.
>
>
>    
Events are atomic in the sense that they are not composed of other events.

Ultimately, it is a question of modelling. If you care about describing

gg shells are broken, egg is stirred, egg mix is slowly added to the cake batter, while the batter is mixed continuously etc

prov-dm tells you to do it with activities.

If you don't care about that level of detail, but just want to say that 
the eggs are used in baking a cake, you just rely on a single event.

Luc

>
>    

-- 
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 Thursday, 8 December 2011 11:31:20 UTC