ProvDM Proposal for restructuring

From Provenance WG Wiki
Revision as of 10:11, 12 May 2012 by Gklyne (Talk | contribs)

Jump to: navigation, search

The master copy of this document is currently maintained offline - please do not edit the document here (for now)

Section references are to http://www.w3.org/TR/2012/WD-prov-dm-20120503/, indicating the intended source of material.

Core

Abstract

Provenance describes the entities, people and activities involved in producing a piece of data or thing, from which the ancestry of an artifact can be discovered in the form of a /provenance trace/, together with information about agencies involved in the various steps leading to that entities existence. The core provenance data model describes elements that comprise the essential structural core of provenance information. This core structure may be enhanced with additional elements, some of which are described in a companion extensions specification, that represent more detailed information about the specific nature of that ancestry.

Introduction

Structure of this Document

Notational Conventions

Preliminaries

Use of PROV-N

Identifiers, namespaces and qualified names

(from 4.7.1, 4.7.2, 4.7.3)

Attributes and values

concept, not specifics

(part from 4.7.4, 4.7.5)

Extension

(refinement)

(from 5)

Validity

(selected material from 6, mainly a reference to the "constraints" document)

Structure of provenance and provenance traces

(material from 2.5, etc.)

Introduce the core triumvirate, with diagram

Diagram/example of a provenance trace (new, simplified)

Provenance core concepts

Entity

(From 4.1.1)

   entity(id, [attr1=val1, ...])

Activity

(from 4.1.2)

   activity(id, st, et, [attr1=val1, ...])

Agent

(from 4.2.1)

   agent(id, [attr1=val1, ...])

Derivation

(from 4.3.1)

   wasDerivedFrom(id, e2, e1, [a], [g2], [u1], attrs)

Use

(from 4.1.4)

   used(id,a,e,t,attrs)

Generation

(from 4.1.3)

   wasGeneratedBy(id,e,a,t,attrs)

[Invalidation]

(from 4.1.7)

   wasInvalidatedBy(id,e,a,t,attrs)

@@I've tentatively included this here as it's a natural counterpart to Generation.

Attribution

(from 4.2.2)

   wasAttributedTo(id,e,ag,attrs)

Association

(from 4.2.3)

   wasAssociatedWith(id,a,ag,[pl],attrs)

("Direction"? - "wasDirectedBy"?)

Responsibility

(from 4.2.4)

   actedOnBehalfOf(id,ag2,ag1,[a],attrs)

("Delegation"? - "wasDelegatedBy")

[communication]

(from 4.1.8)

   wasInformedBy(id,a2,a1,attrs)

@@Two interactions here I think should have a common structural ancestor ("influencedBy"?)

[startedby]

(from 4.1.9)

   wasStartedByActivity(id, a2, a1, attrs)

@@Two interactions here I think should have a common structural ancestor ("influencedBy"?)

Attributes

(from 4.7.4, et seq)

prov:label

prov:location

prov:role

prov:type

prov:value

Extensions

Abstract

The core provenance data model constructs describe the essential structures of a provenance trace, but to not provide specific information about the relationships between entities, activities and agents. Provenance extension constructs may be used to augment or refine the core constructs to provide such additional information about relationships that are commonly encountered in information and data processing activities.

Introduction

Structure of this Document

Notational Conventions

Illustration of PROV-DM by an Example

(From 3, et seq.)

(This could be the current example material, with a simplified example used in the core spec)

The Authors View

The Process View

Attribution of Provenance

Types and Relations

Entities

Specialization

(from 4.4.1)

Alternate

(from 4.4.2)

Activities

Start

(from 4.1.5)

End

(from 4.1.6)

[Communication]

(from 4.1.8, if not in core)

[Start by Activity]

(from 4.1.9, if not in core)

Derivations

Revision

(from 4.3.2)

Quotation

(from 4.3.3)

Original Source

(from 4.3.4)

Trace

(from 4.3.5)


Collections

(From 4.5 et seq)

Collection

Dictionary

Insertion

Removal

Membership

Annotations

(from 4.6 et seq)

Note

(from 4.6.1)

Annotation

(from 4.6.2)