PROV-DM Simplication Reviewer Feedback...

My apologies for being so late on providing reviewer feedback.

Overall I enjoyed the PROV-DM document, I felt that the authors have
done an incredible job helping readers easily relate concepts in the
data model.  Here are my comments and suggestions.

Eric

~~~

Introduction

I agreed with the discussion thread on changes to the introduction
that introduced the purpose of the data model to describe provenance
in natural language.

Section 2.3 – I have mixed feelings about bringing out these concepts,
they don’t tie into the example and collections isn’t mentioned again
until section 5.8.  While they are important perhaps could this
section be left out of section 2?

Section 3 Example

Prior to the auditor example could an ultra simple example debuting an
agent, process and entity something like “w3:Consortium publishes a
technical report”?

I’m wondering if the detailed auditor provenance example could be
introduced first in a human readable story format prior to the
bulleted list that highlights the specific provenance related
concepts.

In the example  use of the somewhat cryptic working draft names
“tr:WD-prov-dm-20111215” “tr:WD-prov-dm-20111018”  is a bit difficult
to following because I found myself mentally parsing the document
names to keep track of the different documents.   While this might be
less realistic something like model-rev1.html, model-rev2.html might
illustrate the same ideas.

I am wondering if it might be more intuitive if the provenance graphic
illustration preceeded the PROV-ASN notation.  It provides a graphic
that a person can study as they study the PROV-DM assertions in
PROV-ASN notation.

The graphic illustration seems to capture all the examples of
provenance from the bulleted list while the PROV-DM assertions in
PROV-ASN seem to be either incomplete (there isn’t a one to one
correspondence to follow from the example to the PROV-DM assertions.

3.2  Great job bringing in the concept of viewing other perspectives
on the same example.

4.2  Activity names in the table need updating.

4.3.3.5  prov:location – Could we change the wording slightly to say
that Location is loosely based on an ISO 19112 but can also refer to
non-geographic places such as a directory or row/column?  The specific
definition from ISO19112 is location:
identifiable geographic place  EXAMPLE “Eiffel Tower”, “Madrid”, “California””

Received on Thursday, 23 February 2012 08:33:20 UTC