AlignmentWithDM

This page is used by the PROV-O.HTML team to align the content and style of provo.html with DM4.

We use this page to communicate with the DM editors to avoid raising too many issues that are only specific to a small group of people.

The alignment is based on the prov-o.html structure http://aquarius.tw.rpi.edu/prov-wg/prov-o retrieved at 16:01pm March 16, and the DM4 http://dvcs.w3.org/hg/prov/raw-file/default/model/releases/WD-prov-dm-20120309/prov-dm.html retrieved at the same time. I have local archives upon request.

In prov-o.html we propose to categorize terms of prov-o into three categories:

• simple
• qualified
• additional

We have a clear notion that qualified terms are used to express n-ary relationships. We are seeking for a clear cut between simple and additional. We are aware of the problem DM faced with respect to core and common and that the DM5 is going to adopt a component based structure. But our original goal of proposing simple is to present a set of terms that are perceived to be most likely used by people to describe simple provenance. But not everybody seems to agree and we are running into a danger of diverting people from the main goal of the charter, which is an interchange language, not a description language.

IMO, agreeing on the set of terms to go into the 1st category of terms in prov-o.html is crucial towards providing a definite structure to work with in the next week. So here are the list of discussion points summarized.

I used the first diagrams in both documents to stimulate the discussion. The first diagram in DM aims to give an overview of the model. And the first diagram in provo.html aims to give an overview of all the simple classes and properties.

What provo.html has but DM does not

• Person, Organization, SoftwareAgent
• wasAttributedTo, startedAtTime, endedByTime, wasInformedBy, tracedTo

What DM has but provo.html does not

• wasStartedBy
• wasEndedBy

PROPOSALS

PROPOSAL 1: Remove all subclasses of Agent in prov-o.html diagram

Remove all subclasses of Agent in the diagram of prov-o.html. We have agreed to remove them from the category of simple terms, but we haven't updated the diagram yet.

PROPOSAL 2: Add temporal properties in the DM diagram

Why are all temporal properties omitted in the DM diagram? By this we refer to startedAtTime, endedByTime.

• it a good suggestion that DM should overview Time attributes. Where? To be discussed with Paolo. (note that DMoverview section is to chane in light of the components)

PROPOSAL 3: Remove tracedTo in prov-o.html diagram and simple

Can it be removed from both the simple category and the first diagram? I am simply proposing this because we have wasDerivedFrom, and we don't need another one.

PROPOSAL 4: Remove wasInformedBy in prov-o.html diagram and simple

If we don't include wasStartedByActivity, I don't see why we include this one here. No matter what we decide I think we should be consistent.

• consistency, both included in simple. (lLM)
• separate discussion diagram vs simple. In prov-dm we will have multiple diagram, one for each component.

PROPOSAL 5: Remove prov:generated from simple category

I think prov:generated is a typo, and should also be removed. Is it right?

• isn't this a relic of the time when there was only entity involvement, and hence you needed the inverse property for generation?

PROPOSAL 6: Remove other classes and properties from simple category

They are not in the diagram of prov-o.html but in the simple category (http://aquarius.tw.rpi.edu/prov-wg/prov-o#prov-o-at-a-glance)

They include: prov:Location, prov:Note, prov:hadLocation, and prov:hasAnnotation.

They are simple, but really not the main thing our DM is about, is it?

ISSUES

ISSUE 1: Not sure what to do with wasAttributedTo in prov-o.html

It's not in the first diagram in DM. Why should it be otherwise in prov-o.html?

• should probably be added to DM overview

ISSUE 2: Align the style of the diagrams in both docs

It would be nice to have the same/similar style for both diagrams. The diagrams in prov-o.html were drawn using powerpoint, in a style similar to those in opmv. How do we want to proceed?

• not a requirement. DM goes for UML diagrams.

ISSUE 2: The purpose of simple

Hopefully by addressing and discussing the above issues we could have a clearer idea about simple.

Jun believes that they should contain the terms that are to be most frequently used by provenance data publishers. Tim believes that they refer to terms that simple as versus those being complex. What do others think and how do we proceed?

• Luc votes for simple / qualified split (though he does not like these names)
• you need a diagram that shows how time is attached to qualified relations.