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

Re: PROV-ISSUE-148 (WasScheduledAfter): wasScheduledAfter definition is difficult to understand

From: Timothy Lebo <lebot@rpi.edu>
Date: Mon, 21 Nov 2011 11:02:42 -0500
Message-Id: <356E48A2-B073-46B8-950B-1BD21F6D6D44@rpi.edu>
To: Provenance Working Group WG <public-prov-wg@w3.org>

On Nov 10, 2011, at 10:47 AM, Provenance Working Group Issue Tracker wrote:

> PROV-ISSUE-148 (WasScheduledAfter): wasScheduledAfter definition is difficult to understand
> http://www.w3.org/2011/prov/track/issues/148
> Raised by: Khalid Belhajjame
> On product: 
> Hi, 
> Trying to model wasScheduledAfter in PROV-O, I found it difficult to understand what wasScheduledAfter means.

It appears to be a very special construct to handle a particular workflow use case.
I'm not sure it will be easily applied to other situations. Thus, I'm rather concerned that it is part of the DM. I think that the "followed" relationship would be much more broadly applicable, and the "wasScheduledAfter" should be an extension that elaborates "followed".

> According to the definition: "Given two activity records identified by pe1 and pe2, the record wasScheduledAfter(pe2,pe1) holds, if and only if there are two entity records identified by e1 and e2, such that wasControlledBy(pe1,e1,qualifier(prov:role="end")) and wasControlledBy(pe2,e2,qualifier(prov:role="start")) and wasDerivedFrom(e2,e1)."
> There are three issues: 
> 1- Why does the agent e2 needs to be derived from the agent e1?

perhaps wasComplementOf would more appropriately model the relationship between the agent ending and the agent starting the processes. (But them we lose the temporal constraint)

> 2- Can an agent be derived from another? (This second issue is secondary).

wasDerivedFrom talks about the _records_ of an agent, not the agents themselves.

> 3- There is an assumption in the definition that the activity pe1 needs to be explicitly terminated by an agent?

That isn't an assumption, it is clearly part of the definition.

> I guess there are cases, the activity will terminates without an agent intervention, and will be followed by the execution of other activity. According to the above definition, in those cases, we will not be able to use wasScheduledAfter.

I agree, we will NOT be able use wasScheduledAfter in those situations (which I would like to be able to model).

So wasScheduledAfter is a specialization of "followed" - which is the more relaxed form that you and I are looking for, I think.


> khalid
Received on Monday, 21 November 2011 16:03:14 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:51:04 UTC