Review of OWL 2 Reference Card (second version of 1 October 2008)

I have three major problems with this document in its current form:

1/ This document is a quick reference guide to the vocabulary of RDF, RDFS,
and OWL, not to the language of RDF, nor to the language of OWL.  This
needs to be reflected in the titles and headings of the document.  

2/ There are many questionable bits of the document, such as the parts
having to do with restrictions.

3/ There are many aspects of OWL that do not fit into the format of the
card.  I do not know whether it is possible to fix the document at all.
One possible way to proceed would be to remove quite a bit of the
content and just list the vocabulary, appropriately grouped.  The other
ways that I see would require a complete rethink of the document.


There are a few minor problems as well:

1/ Some of the headings have no purpose.  Others are grammatically
incorrect. 

2/ The formatting of the document needs to be fixed up in a number of
ways.  In particular, the indentation is inconsistent.



I suggest the following possible minimal change to the first third or so
of the card.  However, there are quite a few areas even in this section
that are problematic and as stated above I am not sure that a useful
document of this form can be produced.

OWL 2 Vocabulary Card

1. RDF/RDFS Vocabulary
-> Remove "RDF Node"
rdfs:Resource	the "universe"
  rdf:type	[...]	membership
  rdfs:label	[...]	human-readable label
  rdfs:comment  [...]	comments
  rdfs:seeAlso	[...]	???
  rdfs:isDefinedBy [...] ???
  rdf:value     [...]   ???
rdfs:Literal	literal values
rdf:XMLLiteral	XML document fragments
-> Remove "Class"
rdfs:Class	RDF classes
  rdfs:subClassOf [...] subclass
-> Remove "Property"
rdf:Property	properties
  rdfs:subPropertyOf [...] subproperty
  rdfs:domain	[...]	domain
  rdfs:range	[...]	range
-> Remove "Containers"
-> Probably best to remove the container vocabulary as it is severely broken.
rdfs:Container	???
  rdfs:member	[...]	container membership
  rdf:_1,...	[...]	container membership
rdf:Alt			...
rdf:Bag			...
rdf:Seq			...
rdfs:ContainerMembershipProperty	???
-> Remove "List"
rdf:List		RDF lists
  rdf:first	[...]	first element of a list
  rdf:rest	[...]	tail of a list
rdf:nil			the empty list
-> Remove "Datatype"
rdfs:Datatype		Datatypes
-> Remove "RDF Reification"
rdfs:Statement		RDF Statements
  rdfs:subject	[...]	the subject of a statement
  rdfs:predicate [...]	the predicate of a statement
  rdfs:object	[...]	the object of a statement

2. OWL Vocabulary
...
-> Remove "Language Elements"
-> Remove "Classes, Datatype and Restriction"
owl:Class		OWL Classes
  owl:intersectionOf	[owl:Class->rdf:List]	Class intersection
  owl:unionOf		[owl:Class->rdf:List]	Class union
  owl:complementOf	[owl:Class->owl:Class]	Class complement

rdfs:Datatype
  owl:datatypeComplementOf [...]	Datatype complement
  owl:withRestrictions	???		???

  owl:oneOf		[owl:Class|rdfs:Datatype->rdf:List]	Sets
  
owl:Restriction		Classes defined by restriction
->	There is quite a bit of stuff here that doesn't fit into the
->	format of the card at all.  I've made a stab at what I think is
->	the least bad thing to do.
  owl:onClass		[owl:Restriction->owl:Class] 					??
  owl:onDataRange	[owl:Restriction->rdfs:Datatype]				??
  owl:onProperties	[owl:Restriction->rdf:List]					??

  owl:allValuesFrom	[...]	??
  owl:someValuesFrom	[...]	??
  owl:hasValue		[...]	??

  owl:cardinality	[...]	??
  owl:maxCardinality	[...]	??
  owl:minCardinality	[...]	??

  owl:qualifiedCardinality	[...]	??
  owl:maxQualifiedCardinality	[...]	??
  owl:minQualifiedCardinality	[...]	??

owl:SelfRestriction	Classes defined by self restriction

  owl:onProperty	[owl:Restriction|owl:SelfRestriction->owl:ObjectProperty|owl:DatatypeProperty]	??

-> remove "(Special classes)"

owl:Thing			OWL individuals
owl:Nothing			Empty class

-> remove "Properties"

owl:DatatypeProperty		Properties with literals as values
owl:ObjectProperty		Properties with individuals as values
owl:AnnotationProperty		Annotation properties

-> remove "(Special properties")

-> And so on, and so forth.



Peter F. Patel-Schneider

Received on Wednesday, 1 October 2008 20:40:32 UTC