W3C home > Mailing lists > Public > public-prov-wg@w3.org > October 2012

Re: PROV-ISSUE-462 (entity-definition-precision): Definition o entity may be too liberal [prov-dm]

From: Paul Groth <p.t.groth@vu.nl>
Date: Mon, 22 Oct 2012 20:26:44 +0200
Message-ID: <CAJCyKRoK6KU2QX4XrbRRfZ_Oe-BcRSgO18-DSwkyLPrJaKUUkA@mail.gmail.com>
To: Luc Moreau <l.moreau@ecs.soton.ac.uk>
Cc: "public-prov-wg@w3.org" <public-prov-wg@w3.org>
Hi Luc,

This is quite a nice clarification. I wonder if there's somewhere we can
use this without adding it to the document? I think the definitions stand
on their own as they now stand.

==Off topic==
In general, there's a "philosophy" that's never been really stated
somewhere that drops out of these clarifications that I think the working
group shares but may not be articulated concisely in a single document. I
think (some) of the key points of the philosophy are:

1) Scruffy ---> Proper
2) Identify the fixed bits your talking about
3) There's multiple kinds of provenance descriptions, we provide a
substrate for all
4) PROV is extensible (it's a substrate)

Maybe this should go in the overview document?


On Mon, Oct 22, 2012 at 11:50 AM, Luc Moreau <l.moreau@ecs.soton.ac.uk>wrote:

>  Dear all,
> I have drafted a response to the following issue. See
> http://www.w3.org/2011/prov/wiki/ResponsesToPublicComments#ISSUE-462_.28Definition_of_Entity.29
>  I will implement the changes once I have a confirmation
> the group is happy with them, and they satisfactorily address the issue.
>  ISSUE-462 (Definition of Entity)
>    - Original email:
>    http://lists.w3.org/Archives/Public/public-prov-comments/2012Jul/0009.html
>    - Tracker: http://www.w3.org/2011/prov/track/issues/462
>    - Group Response:
>       - The term 'entity' is intentionally defined in a liberal manner to
>       avoid restricting users expressivity. Obviously, it shouldn't be too
>       liberal, otherwise it would be all encompassing, without clear semantics.
>       - The term 'entity' (and associated notions such as 'alternate',
>       'specialization') have been the subject of intense debate by the Working
>       Group, and the definition reflects the compromise reached by the Working
>       Group.
>       - The term 'aspect' is not used here with a technical meaning and
>       should be understood with its dictionary meaning 'A particular part or
>       feature of something'.
>       - PROV-Constraints, in its rationale section, expands on the notion
>       of entity.
>       - While an object/thing may change over time, an entity fixes some
>       aspects of that thing for a period of time (in between its generation and
>       invalidation). As opposed to other models of provenance (such as OPM), an
>       entity is not an absolute state snapshot. Instead, it is a kind of partial
>       state, just fixing some aspects. The rationale for this design decision is
>       that it is quite challenging to find absolute state snapshots that do not
>       change: the location of a file on a cloud changes, the footer of this Web
>       page changes (as more people access it), etc. Hence, by allowing *
>       some* aspects (as opposed to all) to be fixed, the PROV concept of
>       'entity' is easy to use.
>       - We distinguish an 'aspect' from an 'attribute'. An
>       attribute-value pair represents additional information about an entity (or
>       activity, agent, usage, etc). In the case of an entity, attribute-value
>       pairs provide descriptions of fixed aspects. So, the term 'aspect' refers
>       to properties of the thing, whereas the term 'attribute' refers to its
>       description in PROV.
>       - PROV does *NOT* assume that all fixed aspects are described by
>       attribute-value pairs. So, there may be some fixed aspects that have not
>       been described. Obviously, without description, it's difficult to query or
>       search over them.
>       - According to PROV Constraint key-object (constraint 23), an
>       entity has a set of attributes given by taking the union of all the
>       attributes found in all descriptions of that entity. In other words, PROV
>       does not allow for different attribute-value pairs to hold in different
>       intervals for a given entity.
>       - The attribute-value pairs of an entity provide information for
>       some of the fixed aspects of an entity.
>          - *This point may not have been clear, and requires text
>          modification*. (see below)
>       - A specific attribute of an entity is its identity. It is also
>       assumed that it holds for the duration of the entity lifetime.
>          - *This point may not have been clear, and requires text
>          modification*. (see below)
>    - References:
>       - PROV constraints rationale:
>       http://www.w3.org/TR/2012/WD-prov-constraints-20120911/#entities--activities-and-agents
>       - entity/specialization/alternate definitions:
>       http://www.w3.org/2011/prov/wiki/SpecializationAlternateDefinitions
>       - Resolution on entity/specialization/alternate:
>       http://www.w3.org/2011/prov/meeting/2012-05-03#resolution_2
>       - Key Constraints definition:
>       http://www.w3.org/TR/2012/WD-prov-constraints-20120911/#dfn-key-constraints
>       - Key-Object constraint 23:
>       http://www.w3.org/TR/2012/WD-prov-constraints-20120911/#key-object
>    - Proposed Changes to the document:
>       - http://www.w3.org/TR/prov-dm/#entity.attributes: instead of
>       "representing additional information about this entity." write
>       "representing additional information about the fixed aspects of this
>       entity."
>       - http://www.w3.org/TR/prov-dm/#term-identifier: add the following.
>          - Entity, Activity, and Agent have a mandatory identifier. Two
>          entities (resp. activities, agents) are equal if they have the same
>          identifier.
>          - Generation, Usage, Communication, Start, End, Invalidation,
>          Derivation, Attribution, Association, Delegation, Influence have an
>          optional identifier. Two generations (resp. usages, communications, etc.)
>          are equal if they have the same identifier.
> Luc
> On 07/25/2012 08:16 AM, Provenance Working Group Issue Tracker wrote:
> PROV-ISSUE-462 (entity-definition-precision): Definition o entity may be too liberal [prov-dm]
> http://www.w3.org/2011/prov/track/issues/462
> Raised by: Paul Groth
> On product: prov-dm
> This is the issue for http://lists.w3.org/Archives/Public/public-prov-comments/2012Jul/0009.html
> from Jacco van Ossenbruggen
> --
> 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

Dr. Paul Groth (p.t.groth@vu.nl)
Assistant Professor
- Knowledge Representation & Reasoning Group |
  Artificial Intelligence Section | Department of Computer Science
- The Network Institute
VU University Amsterdam
Received on Monday, 22 October 2012 18:27:13 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:58:20 UTC