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

Re: PROV-ISSUE-509: Data Model Figure 5 [prov-dm]

From: Luc Moreau <l.moreau@ecs.soton.ac.uk>
Date: Wed, 26 Sep 2012 11:07:53 +0100
Message-ID: <EMEW3|c20ab1c95c69d712462d8fd79e5bce89o8PB7s08l.moreau|ecs.soton.ac.uk|5062D3F9.6030704@ecs.soton.ac.uk>
To: public-prov-wg@w3.org
Dear all,

I have drafted a response to ISSUE-509 on the wiki
Feedback appreciated,


      ISSUE-509 (AttributesInUML)

    * Original email:
    * Tracker: http://www.w3.org/2011/prov/track/issues/509
    * Group Response
          o First, let us note the non normative nature of the UML
            diagrams. They are here to inform readers, and convey the
            intuition of the data model
          o The UML actually represent all the information present in
            relations such as WasStartedBy.
                + PROV Id and PROV attributes are explicitly listed as
                  UML attributes in the association class
                + The started activity and the trigger entity are source
                  and destination of the association edge
                + The starter activity is present with the starter edge
                + Time is also present though the time edge
          o With UML diagrams, we can take a full object oriented view
            or a more relational view of the data model. The former
            lists all attributes, whereas the latter highlights the
            relations. We opted for the latter approach.
          o Hence, what the UML diagram does not explicit represent is
            the actual names of all attributes of a relation. That is
            covered by the normative text.
          o It is correct that Time is a primitive datatype, and marked
            as such. Given the important of time and events in the
            model, it is considered pedagogical to keep it in Figure 5.
            We note that Figure 1, the much simplified version, doesn't
            show it.
          o Finally, it's correct that we use names such as Start, but
            the UML diagram contains relation label WasStartedBy. This
            has now been fixed for all introductory paragraphs.
    * References:
          o UML informative nature:
    * Implemented changes:
    * Original author's acknowledgement:


On 10/09/2012 09:39, Provenance Working Group Issue Tracker wrote:
> PROV-ISSUE-509: Data Model Figure 5   [prov-dm]
> http://www.w3.org/2011/prov/track/issues/509
> Raised by: Luc Moreau
> On product: prov-dm
> http://www.w3.org/2011/prov/wiki/LC_Feedback#Data_Model_Figure_5
> ISSUE-463
> The figure is a little cluttered. Since Time is a primitive datatype and not an object, it could be removed for clarity. Other datatypes are not shown.
> The headers of the supporting text in section 5.1 are different than the labels in the figure, so the reader must use table 5 as a cross reference when going between them. Some translations are straightforward, but for readers that are new to PROV it is not obvious that wasInformedBy means Communication (for example).
> When modeling complex relationships that have mandatory and optional attributes, class associations are a good approach to use. The current version of the spec does not fully model the relationships, however. Each of the gray classes in figure 5 could be modeled as a complete class that follows the requirements of the relationship. For example, the class wasGeneratedBy currently shows only two attributes, although the spec for Generation lists five (four of which are optional, and one of which can be zero to many). All of this information can be represented very concisely in the diagram, which would make it easier for a reader to get a complete view of the spec graphically.

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 Wednesday, 26 September 2012 10:08:19 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:51:21 UTC