W3C home > Mailing lists > Public > www-webont-wg@w3.org > December 2002

Re: Feature Synopsis Review

From: Deborah McGuinness <dlm@ksl.Stanford.EDU>
Date: Tue, 31 Dec 2002 19:00:57 -0800
Message-ID: <3E1259E9.AECEE0@ksl.stanford.edu>
To: Jeremy Carroll <jjc@hpl.hp.com>
CC: www-webont-wg@w3.org, dlm@ksl.Stanford.EDU

Thanks for the comments.
The response to these comments also addresses the action item I had to further
incorporate the 3 levels of owl into the feature synopsis document.
This used (and continues to use) the guide document from which to either draw
text or to point to when summarizing for the species of owl.

specific comments below:

Jeremy Carroll wrote:

> This review is limited to those issues indicated in:
> http://lists.w3.org/Archives/Public/www-webont-wg/2002Dec/0320.html
> The feature synopsis is badly misleading/incorrect on these issues.
> In particular the last paragraph of section one incorrectly suggests that OWL
> DL and OWL Full differ semantically rather than syntactically.

The last paragraph attempted to summarize the guide section that was too long for
this document which says:
"OWL DL includes the complete OWL vocabulary, interpreted under a number of
simple constraints.... " and
"OWL Full includes the complete OWL vocabulary, interpreted more broadly than in
OWL DL, with the freedom provided by RDF.... "

into the summary:
There are two interpretations for the full OWL vocabulary - one used in OWL DL
that is more restricted and one used for OWL Full that is less restrictive. Since
the vocabularies for OWL DL and OWL Full are identical, this document only
distinguishes between the OWL Lite vocabulary and the full OWL vocabulary. "

are you making a proposal that both the guide and the feature synopsis
incorporate new proposed text?
if so, could you suggest it?

> No awareness of the syntactic restrictions on OWL Lite or OWL DL other than
> the vocabulary is indicated.

there previously was an introduction in 3 - language description of owl lite that
mentioned owl lite has a subset of the full owl langague constructors and has a
few limitations.
were you proposing adding more text here?

> The summary suggests that OWL has two levels OWL Lite and OWL Full, rather
> than three levels ...

did you mean the abstract?
I had originally left the abstract with just the owl lite vocabulary and the full
owl vocabulary and then in the introduction discussed the different
interpretations with the 3 languages.  Since that may have been confusing, i
modified the abstract to explicitly mention all three and to point people to the
guide for more info on the distinctions.

> The header information section 3.6 needs updating (I am not sure which of the
> header stuff is in OWL Lite or OWL DL; maybe there is a need for a new
> section "OWL DL Header Information").

My understanding is that owl lite has the header information in the feature
synopsis document.
I missed whenever we put the others in but I can add the ones in the reference
document including:


        (By the way - i notice that the guide document has backCompatibleWith
        and the reference document has backwardCompatible - lets choose one.)


I dont remember when we decided what level these went into.
My default will be to add them to owl lite but i can easily put them in the owl
dl/full section if that is preferable.

> The document needs another section to indicate the relationship between RDF
> instance data and OWL Lite instance data. Specifically RDF instance data is
> only OWL Lite instance data if the following conditions are met:
> - any class used is explicitly declared as of type owl:Class.
> - no bnode is the object of two or more triples.
> - any datatypes used are explicitly declared as of type rdfs:Datatype
> (such class and datatype declarations must be explicitly in the file; an
> imports is insufficient). Given that syntax is explicitly excluded from the
> scope of this document, it may suffice simply to say that there are
> restrictions.

Is this discussion in the guide or the reference document?  the feature synopsis
should just summarize this and point to it.
I looked but did not find it.  (i did some searches for bnode, instance data,
can someone provide a section to point to?

> The treatment of Dublin Core metadata needs expanding. As is, the document is
> contradicted by the mapping rules. The mapping rules only permit annotations
> being simple triples where the subject is the document and the object is a
> URIref or literal value. Dublin Core metadata also permits the object to be a
> complex structure e.g. represented by a blank node or uri ref with further
> triples defining it.

I had asked for input earlier on what to include here.  I included what I

two thoughts:
1 - the reference doe not include a mention of dublin core at all.  i could drop
this from the short feature synopsis.
2 - dublin core could continue to be mentioned in the feature synopsis.
the guide has:
"One common set of additional tags that could reasonably be included here are the
standard Dublin Core metadata tags. Examples include Title, Creator, Description,
Publisher, and Date. The URI defining the Dublin Core namespace is
'http://purl.org/dc/elements/1.1/'.) "

the feature currently has:
Dublin Core MetaData: Ontologies also have a non-logical component (not yet
specified) that can be used to record authorship, and other non-logical
information associated with an ontology. A suggestion is to associate attributes
from the Dublin Core Metadata standard with the ontology.
(with the link brian provided for dublin core)

Could you propose what should be included that is short?

> A further problem is the discussion of intersectionOf.
> The mapping rules from the OWL Lite abstact syntax use intersectionOf and so
> it is incorrect to assign it only to OWL DL/OWL Full. However, the use in OWL
> Lite is highly restricted.

this is worth a short conversation.  while i wanted intersection in owl lite, in
an effort to add only a small number of non-controversial constructors, it was
left out.
possibly the best solution is to include a pointer to the semantics document to
see the restricted use of intersection in owl lite.

> Overall I agree with Peter that this document should be merged into the
> reference.

I strongly support keeping the document separate and short basically in its
current form.
I strongly oppose merging it into another document.
I have gotten a lot of feedback since it was publically announced in july and
even before of how useful people found it to have a short document that had a
high level discussion.  they did not want the syntax - they just wanted a one (or
two) sentence description and example.

> The exclusion of syntactic considerations from the scope of this
> document is a critical omission for a recommendation that if it is to serve
> any purpose acts to define the OWL Lite/OWL DL/OWL Full levels.
> More minor points
> ==============
> Suggest delete all references to DAML+OIL. (Superfluous).
> In Introduction penultimare para suggest delete sentences 2 and 3 from "The
> goal of OWL Lite .... will flock." (Superfluous).
> The reference to the mailing list at the end of section 3.4 is inappropriate
> in a recommendation in my view.

this was required at the last face to face meeting since we explicitly rejected
the user proposal to change the names.  The group decision was to point to the
message that chris sent to the mailing list.

> The last sentence about previous versions is superfuous - and should be
> deleted.
> Jeremy

 Deborah L. McGuinness
 Knowledge Systems Laboratory
 Gates Computer Science Building, 2A Room 241
 Stanford University, Stanford, CA 94305-9020
 email: dlm@ksl.stanford.edu
 URL: http://ksl.stanford.edu/people/dlm
 (voice) 650 723 9770    (stanford fax) 650 725 5850   (computer fax)  801 705
Received on Tuesday, 31 December 2002 22:01:58 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:56 GMT