W3C home > Mailing lists > Public > public-owl-wg@w3.org > April 2009

Re: Quick Review of New Features and Rationale

From: Christine Golbreich <cgolbrei@gmail.com>
Date: Wed, 29 Apr 2009 17:14:50 +0200
Message-ID: <b0ed1d660904290814m48bf4870w32b46527964eec46@mail.gmail.com>
To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
Cc: W3C OWL Working Group <public-owl-wg@w3.org>, Evan Wallace <ewallace@cme.nist.gov>
Hi Peter,

Thank you for this review.

Quick response :

In general, I agree with your comments. However, it is sometimes
tricky to conciliate with other constraints, and to integrate the new
constructs that have been added recently into the existing
organisation, which had its own "logics". In particular, the two main
concerns are
1) that the document remains understandable by users without supposing
too much background / implicit knowledge. Since we have removed much
content/explanation, it is necessary in my view, to  keep for example
at least some "headers" or structurantion so as to be  minimally
explicit. Only the grammar is short for people who don't know the
meaning before.
2) features F1 to F15 are associated to UCs/Examples, serve as
motivations, and have been built on the same pattern (see intro of
It's why it's a little difficult to insert other content in the
middle, even though it might be more logical to have a different order
from another point of view . I did my best !

I have already implemented the most important. The only point I
disagree is to remove the Note before 2.2.4 because it was specially
added to answer LC1 comments (namely from Frank, and Jeremy).

See below details

2009/4/23 Peter F. Patel-Schneider <pfps@research.bell-labs.com>:
>        Quick Review of NF&R
> I reviewed the version of 23 April 2009.
> General Comments:
> - Links to WG documents should use [[...]], not [...].  However, given
>  that all the links are there, they might not be worth changing now.

will do asap

> - The examples need to be reformatted to match what is done in the other
>  WG documents for multiple syntaxes.

I  have started, see F1 to F4.
I mimic the method used in the Syntax, adding FFS and RDF buttons to
Hide/Show the 2 syntaxes within the Examples.

> - Names in the examples need to be fixed to match the language, i.e.,
>  abbreviated IRIs need to be "qualified".

not sure what you mean by "qualified" ?
do you mean to add  a namespace/prefix e.g.  a:person ? or ":" for
each entity or something else?
I believe that adding a namespace  ":a" everywhere is heavy and does
not bring much, but I can make it if necessary.

> - There are too many sub-...-sections and headers in the document.  To
>  help this a bit and should not be separate sections,
>  nor should there be headers in Section 4.1.2

did my best, will stil try to improve at some places, see above

> - Section 2.7 is internally self-contradictory.  The first sentence says
>  that it doesn't talk about anything that increases expressive power,
>  but 2.7.3 is about boolean combinations of data ranges, which does
>  increase expressive power.  Section 2.7.3 should be moved to be part
>  of Section 2.3, both for this reason and also to move it with the other
>  datatype stuff.


> - Both references sections need to be fixed up.  Section 6 should be
>  ordered.  Many references in Section 7.22 are missing components.
>  Each reference to a wiki page should be examined to see if some better
>  reference can be found.  If not, both the page and the reference
>  should be augmented with as mouch attribution information as possible.

still TBD

> Specific Comments:
> - The abstract should start "This document presents a simple overview"
> - The paragraph before 2.2.4 could be removed.

see above

> - The document needs to include the new DisjointDataProperties axiom.


> - The document needs to include datatype definitions, in Section 2.3.


> - Section 2.5 needs significant work.  I suggest the following changes,
>  some of which fix mistakes in the document:


did not see what was "significant", please have a look.

>  2.5.1 ...
>  ....
>  OWL 2 provides the construct <span class="nonterminal"
>  id="a_AnnotationAssertion">AnnotationAssertion</span> for annotations
>  on ontologies, entities (such as classes or properties),  and
>  anonymous individuals.
>  In the OWL 2 Direct Semantics, these annotations
>  carry no logical import, allowing the direct use of DL reasoners.
>  <grammar>
>  <example>
>  OWL 2 provides the <span class="nonterminal"
>  id="a_Annotation">Annotation</span> constructing for annotating axioms
>  and other annotations.
>  These annotations also
>  carry no logical import
>  in the OWL 2 Direct Semantics.
>  <grammar>
>  <example> - use 'Annotation(rdfs:comment' instead of 'Comment('
>  2.5.2 Axioms for annotation properties
>  Annotation properties can be given domains (<span class="nonterminal"
>  id="a_AnnotationPropertyDomain">AnnotationPropertyDomain</span>) and
>  ranges (<span class="nonterminal"
>  id="a_AnnotationPropertyRange">AnnotationPropertyRange</span>) and
>  participate in an annotation property hierarchy (<span
>  class="nonterminal">SubAnnotationPropertyOf</span>).
>  These constructs
>  have no logical import in the OWL 2 Direct Semantics.
>  <grammar> - add domain and range gramma
>  <example>
> - The discussions of each profile need to be turned into sentences.

will do (TBD)

> Tyopgraphical but Grammatical Erors:

> I fixed a number of typographical and minor grammatical errors.  Quite a
> few remain, however.

Received on Wednesday, 29 April 2009 15:15:32 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:42:12 UTC