Re: new editor's draft of S&AS [multiple restrictions in OWL Lite?]

According to the changes in the Abstract Syntax [1]: 

>> CHANGES from informal meeting - DONE
>> 
>> - OWL Lite restrictions must have exactly one component
>>   - to reduce complexity in recognizing OWL Lite graphs
>>   - changed Abstract Syntax
>> 
>> - OWL DL restrictions must have at least one component
>>   - to fix bug in mapping rules
>>   - changed Abstract Syntax
 
the occurence of restriction components is limited 
exactly one under the "restriction" non-terminal [2]. 
However, multiple restriction components are allowed 
in a class axiom via the "super" non-terminal for 
OWL Lite [3]. 

Isn't this a problem for the Abstract Syntax of OWL Lite? 


> 2.3.1.1. OWL Lite Class Axioms 
> 
> axiom ::= 'Class(' classID ['Deprecated'] modality { annotation } { 
super } ')'
> super ::= classID | restriction


[1] http://lists.w3.org/Archives/Public/www-webont-wg/2003Mar/0064.html
[2] 
http://www-db.research.bell-labs.com/user/pfps/owl/semantics/syntax.html#2.3.1.2
[3] 
http://www-db.research.bell-labs.com/user/pfps/owl/semantics/syntax.html#2.3.1.1


Masahiro Hori, Ph.D. 
Group Leader, Programming Models & Tools, 
IBM Tokyo Research Laboratory
Tel: +81-46-215-4667 / Fax: +81-46-274-4282
Email: horim@jp.ibm.com





"Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
Sent by: www-webont-wg-request@w3.org
2003/03/11 23:39
 
        To:     www-webont-wg@w3.org
        cc: 
        Subject:        new editor's draft of S&AS



There is a new editor's draft of S&AS in the usual place (follow the link
on the WebOnt home page) that has many of the changes discussed at the
informal meeting last week.

peter

PS:  Here are some somewhat idiosyncratic notes on the changes and how 
S&AS
     has changed to accommodate them.  These notes should only be used as 
a
     guide to what has recently changed in S&AS.


                                 Notes on Recent Changes to OWL S&AS


CHANGES from informal meeting - DONE

- OWL Lite restrictions must have exactly one component
  - to reduce complexity in recognizing OWL Lite graphs
  - changed Abstract Syntax

- OWL DL restrictions must have at least one component
  - to fix bug in mapping rules
  - changed Abstract Syntax

- wording change to transitivity vs cardinality side condition
  - changed Abstract Syntax

- required axiom for annotation properties
  - allows annotations
  - changed Abstract Syntax
  - changed Direct Semantics

- allow anonymous individuals as objects of annotations
  - changed Abstract Syntax
  - changed Direct Semantics

- type all URI references and blank nodes
  - including annotation properties and ontologies
  - changed Abstract Syntax

- list excluded vocabulary
  - don't disallow RDF container vocabulary and RDF reification vocabulary
    - need to use the vocabulary consistently
  - changed Mapping
  - *need to change RDFS semantics*

- non-recommended datatypes
  - allow other datatypes as uninterpreted datatypes
  - changed Abstract Syntax
  - changed Direct Semantics

CHANGES for other reasons

- fixes to the proofs
  - added annotations and imports
  - needed some changes to definitions

- cleanup
  - 



CHANGES - TODO

Typing (Mapping rules)
  - add owl:DataRange as well
  - even if only used the object of triple

Orphans
- allowed in DL via n=1 equivalentclasses hack
- add text on this issue in abstract syntax

OntologyProperties *UNRESOLVED*
- resolution - allow them !!

rdf:XMLLiteral *UNRESOLVED*
- proposal 1 - allow rdf:XMLLiteral as uninterpreted datatype
- proposal 2 - allow rdf:XMLLiteral as in RDF

reverse mapping
- wording from Jeremy


Issues

* Bugs in secondary description of OWL DL ontology in triples
  - e.g. - n-ary equivalence and disjointness


POTENTIAL CHANGES

Descriptions
- allow descriptions to share structure
- allows for arbitrary equivalent/different patterns
- investigate whether this can work

Questions
- can we allow arbitrary graphs of unnamed individuals? 
  - NO! - messes up direct semantics

To do
- look into naming data ranges