Review of Fragments Doc for ACTION-110

To dispatch ACTION-110, I performed a detailed review of the current
fragments document.

The benefit of following Achille's detailed review is not having to
repeat his comments [1].  Below are some additional items I found during
review, or particular points I wanted to emphasize.


Abstract:

+ The first two sentences of the abstract appear to be irrelevant to
this document.


Section 3: DL-Lite:

+ I added an editor's note referencing [ISSUE-80] (which addresses
selection of a particular DL-Lite fragment from within the family) in
the style of similar notes in the Syntax doc.

(An aside: what's the intended difference between a reviewer's note and
an editor's note - we still don't have official editors correct?)


Section 3.1: DL-Lite Feature Overview:

+ ObjectPropertyAssertion is omitted from the list of permissible facts


Section 4.1: OWL-R Feature Overview:

+ The vocabulary used here is unnecessarily different from the
vocabulary used to describe similar constraints for the other fragments.
E.g., in section 3.1 "the following constructs can be used to define
subclasses in SubClassOf axioms", but in 4.1 "the following constructs
can be used in the antecedents of class implications"


Section 5: Computational Properties

+ I agree with and second Achille's proposal to include references for
complexity results.


Additionally, I am uncertain about,

Section 2.2.4: EL++ Axioms:

+ Are DisjointDataProperties axioms supposed to be in EL++? (The spec
currently includes them?  I am not that familiar with the literature
here and couldn't quickly find them in the references.  I ask b/c
DisjointObjectProperties are explicitly disallowed.


Regards,
-- 
Mike Smith

Clark & Parsia

[1] http://www.w3.org/mid/OFD7BC8F65.EB99FEFF-ON8525740F.0013B481-8525740F.0014770D@us.ibm.com
[ISSUE-80] http://www.w3.org/2007/OWL/tracker/issues/80

Received on Monday, 17 March 2008 22:04:50 UTC