- From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
- Date: Mon, 17 Mar 2003 12:15:47 -0500 (EST)
- To: www-webont-wg@w3.org
Comments on Overview and Guide
(I'm waiting for the new version of Reference before providing comments.)
- title - I changed to
OWL Web Ontology Language
Semantics and Abstract Syntax
on the suggestion of Jim Hendler. This makes it clear that OWL
is not a misspelling of an acronym. I suggest that we again
revisit the idea of a common way of naming the documents.
Comments on ``Web Ontology Language (OWL): Overview'' dated ``W3C Working
Draft 4 March 2003''.
general comment - I STRONGLY suggest using owl: for all OWL terms.
Reference and Semantics and Abstract Syntax already do so.
Guide mostly does so also.
general comment - I again suggest removing Section 2. In my opinion,
Section 2 is decidedly harmful. Note in particular, the bare
inclusion of intersectionOf in Section 2.1. How will this be
understood by readers? If a language synopsis is needed, it could
be an appendix.
general comment - I remain skeptical as to the utility of this document.
Maybe it is because I do understand OWL, but I find that when
reading this document I end up with many more questions than
answers.
missing - annotations
- name separation in OWL Lite
- ObjectProperty and DatatypeProperty separation in OWL Lite
- oneOf over data values
- references section
title - I suggest
The OWL Web Ontology Language: Overview
abstract - I suggest starting
The OWL Web Ontology Language is designed for use by
applications that need to process the content of
information instead of just presenting information to
humans.
web vs Web - I think that ``Web'' is supposed to be used instead of
``web''.
abstract - I suggest
OWL has three increasingly-expressive sublanguages: OWL
Lite, OWL DL, and OWL Full.
abstract - OWL Semantics and Abstract Syntax
status - I suggest
This overview ... herein are provided to help understand
OWL, but may not ... OWL Semantics and Abstract Syntax.
1. - I suggest
The document describes the OWL Web Ontology Language. OWL
is intended to be used when the information contained in
documents needs to be processed by applications, as opposed
to situations where the content only needs to be presented
to humans. ... is called an ontology.
1.1 - I suggest
The following ... .
- This overview ... .
- The OWL Guide ... .
- The OWL Reference ... .
- The OWL Semantics and Abstract Syntax ... .
1.2 - I suggest
... The first level above RDF required for the Semantic Web is an
ontology language what can formally describe the meaning of
terminology used in Web documents. ... OWL Use Cases and
Requirement provides ... for a Web Ontology Langauge.
OWL has been designed to meet this need for a Web Ontology Language.
OWL is part of the growing stack of W3C recommendations related to
the Semantic Web.
(don't use a dl here, instead use a ul, as in Section 1.3)
1.3 - I suggest
OWL Lite ... constraints, OWL Lite ... and OWL Lite provides ...
OWL DL ... want maximum expressiveness while retaining
computational ...
OWL Full ... unlikely that any reasoning software will be able to
support complete reasoning for every feature of OWL Full.
... users require the more-expressive constructs provided by OWL
DL. ... e.g., ... classes or attaching
... Every OWL (...) document ...
1.4 - I suggest
... features of OWL Lite ... OWL Full. OWL DL and ... .
2 - I suggest
This section provides ....
2 - See comment on using owl:
2.1 - I suggest
The list ... is given below.
2.1 - The RDF Schema Features list is confusing to me. What does
Individual ... have to do with RDF Schema?
2.1 - Something needs to be added on datatypes.
2.2 - I suggest that the HTML source be modified to give much less
spacing
2.2 - I suggest
The list ... is
3 - I suggest
... has more limitations on the use of the features than OWL DL or
OWL Full. In OWL Lite classes ... (cannot be arbitrary
expressions), ... are also only allowed between named classes,
... Similarly, restrictions
3.1 - The links under SubClassOf and AllValuesFrom are not appropriate.
3.1 - I suggest
owl:Class: ... owl:Thing ... a superclass of all OWL classes.
3.1 - I suggest
rdfs:subClassOf: ... subclass of another class. ... if an individual is a
Person, then it is also a Mammal.
3.1 - I suggest
rdf:Property: Properties can be used to state relationships between
individuals or from individuals to data values. ... include
hasChild, .... The first three can be used to relate .... The last
(hasAge) can be used to relate ...
3.1 - I suggest
rdfs:subPropertyOf: .... that if an individual is related to
another ..., then it is also related to the other by ....
3.1 - I suggest
rdfs:domain: A domain of a property limits the individuals to
which the property can be applied. If a property relates an
individual to another, and the property has a class as one of its
domains, then the individual must belong to the class.
3.1 - I suggest a similar change for rdfs:range
3.2 - I suggest
owl:equivalentClass: Two classes may be stated to be equivalent.
Equivalent classes have the same instances. ...
3.2 - I suggest a similar change for equivalentProperty
3.3 - I suggest not using SSN as an inverse functional property example, as
it could be misconstrued as implying that OWL Lite allows inverse
functional on datatype properties.
3.4 - I suggest merging 3.4 and 3.5
3.4 - I suggest
OWL Lite allows restrictions to be placed on how properties can be
used by instances of a class. The first two kinds of restrictions
limit which values can be used, the last three kinds limit how may
values can be used.
3.4 and 3.5 - I suggest rewriting all five points to fit with the
introductory wording above.
3.7 - I suggest something like:
OWL uses the RDF mechanisms for data values. See the OWL Guide for
more information.
3.8 - I suggest
OWL Lite supports notions of ontology inclusion and relationships
and attaching information to ontologies. See OWL Reference for
details and OWL Guide for examples.
5. - I suggest
... Web ...
5. - The acknowledgments should go in an Acknowledgments section.
Comments on ``Web Ontology Language (OWL) Guide Version 1.0'' dated ``W3C Working
Draft 9 March 2003''.
general question - Why are Overview Section 1.3 and Guide Section 1.1 so
similar?
general comment - It is OWL Semantics and Abstract Syntax, not OWL Abstract
Syntax and Semantics.
web vs Web - I think that ``Web'' is supposed to be used instead of
``web''.
title - I suggest
The OWL Web Ontology Language: Guide
abstract - I suggest
The OWL Web Ontology Language is intended ....
missing - oneOf on datavalues
1 - I suggest no paragraph break before ``To support ...''.
1 - I suggest
The OWL Web Ontology Language is a language ...
1 - I suggest
An OWL ontology may include information about classes, properties,
and their instances.
1.1 - I suggest
... while OWL Lite supports ... and provide a quick ...
2 - I suggest
OWL is ... content.
OWL collects related information into ontologies.
An OWL ontology is mostly a collection of information about
classes, properties, and their instances.
As OWL is part of the Semantic Web, and the Semantic Web is
inherently distributed, OWL must allow for information to be
gathered from distributed sources. This is partly done by allowing
ontologies to be related, including explicitly importing
information from other ontologies.
The other way that OWL deals with distributed information is by
working within an open world framework. In OWL, information about
a resource is generally not assumed to be complete. While a class
or individual can be introduced in a particular ontology and
information about the class or individual can be given there,
information about the class or individual can also be given in
other places. New information can only extend
previously-encountered information, never override it, making OWL a
monotonic formalism.
New information can contradict previously-encountered information,
but even this situation does not result in the
previously-encoutered information being overriden, instead
resulting in a formal contradiction. The possibility of such
contradictions is something that a designer of an ontology needs to
take into consideration. Is is expected that tool support will
help detect such cases.
2.1 - I suggest
Before we can use a set of terms in OWL, ....
2.1 - I suggest
OWL depends on constructs defined by RDF, RDFS, and XML Schema
datatypes. ...
2.2 - I suggest
The owl:Ontology element ...
2.2 - I think that it is ``xml:base''.
2.2 - I suggest
... mechanisms. Namespace declarations provide .... indicate the
intention to include ...
2.3 - It is ``rdf:about''.
2.3 - I suggest
Properties that are used as general annotations ...
2.3 - I suggest removing the paragraph on </owl:Ontology> this closing tag
is given above.
2.3 - I suggest
is ultimately closed by
</rdf:RDF>
which closes the namespace declaration shown above.
3 - It is generally not a good idea to start a section with a subsection
heading. I suggest starting this section as
3. Basic Ontology Information
Much of the information in an OWL ontology concerns classes, properties,
instances of classes, and relationships between these instances.
Some notions of ontology only permits the first two kinds of
information allowed in an OWL ontology, but OWL also allows
information to be specified about particular individuals.
3.1.1 - I suggest
3.1.1 Simple Named Classes
owl:Class, rdfs:subClassOf
3.1.1 - I suggest
... In particular, we will have more to say about Winery later.
The syntax rdf:ID="Region" is used to introduce a name. This is
...
3.1.1 - I suggest
It is possible to refer to
3.1.1 - I suggest
... using its full URI, here ...
3.1.1 - I suggest
The simplest taxonomic constructor for classes is rdfs:subClassOf.
This relationship relates a more-specific class to a more-general
class. If X is a subclass of Y, then every instance of X is also
an instance of Y. The rdfs:subClassOf relationship is transitive.
...
3.1.1 - I suggest
Information about a class usually also includes restrictions on
instances of the class. So far we have ...
3.1.2 - I suggest
3.1.2 Individuals
3.1.2 - I suggest
Note that the following is identical to the introduction above.
3.1.2 - I suggest no paragraph break before ``Second, ...'.
3.1.2 - I suggest
In order to have available a few more classes for the examples in
the next section, we introduce ...
3.1.2 - I think that ``underly'' should be ``underlie''.
3.2 - I suggest
3.2 Simple Properties
This world of classes and individuals ....
3.2.1 Defining Properties
owl:ObjectProperty, owl:DatatypeProperty, rdfs:subPropertyOf,
rdfs:domain, rdfs:range
3.2.1 - I suggest removing the sentence before ``The property
madeFromGrape...''.
3.2.1 - I suggest removing the sentence after ``... made from at least one
WineGrape.''
3.2.1 - Highlighting may not show up on all renderings of the document. I
suggest using instead
The restriction
... (include the restriction explicitly)
defines ...
3.2.1 - Restrictions are allowed in OWL Lite, and thus should not be tagged
as OWL DL.
3.2.1 - I suggest
Including this restriction
3.2.1 - I suggest avoiding the word ``cliche'' to describe simple
restrictions. I view it as much better to reserve ``cliche'' for
larger constructs that could be written in several ways.
3.2.1 - I suggest
We can now describe the class of Vintages, ...
3.2.2 - I suggest
make use of many of the built-in XML Schema datatypes. References
to these datatypes ... suitable ....
3.2.2 - I suggest
... caveats described in OWL Semantics and Abstract Syntax.
3.2.2 - I suggest moving the tennisGameScore example to the end of 3.2.2.
3.2.2 - I suggest
... We introduce the hasVintageYear ...
3.2.3 - I suggest
First we describe Region and Winery individuals, and then we
introduce our first wine, ....
3.2.3 - I suggest
Below we describe an instance of VintageYear ...
3.3 - I suggest
to further specify properties and how they can be used.
3.3 - I suggest not using ``tagged''. A much better phrase would be
``specified as''.
3.4 - I suggest having only one paragraph here.
3.4.1 - I suggest
... restriction requires that ...
3.4.2 - I suggest
... we specify Vintage ...
3.4.2 - I suggest
We specified hasVintageYear as a functional property.
3.4.2 - I suggest removing the last sentence, as it can be confused with
ranges like integers between 1800 and 1900.
3.4.3 - I suggest
hasValue allows us to specify ...
4.1 - I suggest
4.1 Equivalence between Classes and Properties
owl:equivalentClass, owl:equivalentProperty
To tie together a set of ontologies it is frequently ...
4.1 - I suggest
... owl:equivalentClass ..
4.1 - I suggest
The property owl:equivalentClass ... have precisely the same
instances. Note that in OWL DL, .... ... we can use
owl:sameIndividualAs ...
4.1 - I suggest
... two independently developed ontologies, ...
owl:equivalentClass ...
4.1 - I suggest
... rdfs:subClassOf ... owl:equivalentClass ...
... owl:equivalentClass ...
(etc, in this section and later in the document)
4.2 - I suggest
4.2. Identity between Individuals
owl:sameIndividualAs, owl:sameAs
4.3 - I suggest
4.3. Different Individuals
owl:differentFrom, owl:AllDifferent
5 - I suggest
These are called owl:....
5.1 - I suggest regularizing the subsection headings like
5.1.1 Intersection [some uses OWL DL]
owl:intersectionOf
5.1.2 Union [OWL DL]
owl:unionOf
5.1.3 Complement [OWL DL]
owl:complementOf
5.1.1 - All the examples are in OWL Lite. I suggest adding wording that
talks about the difference, essentially saying that some uses of
intersection are in OWL Lite. This will also affect the wording in
5.1.2 and the tagging of Section 5.
5.2 - I suggest
OWL provides the means to specify a class via
5.2 - I find the last example confusing. I think that a complete oneOf
should be used.
6 - I suggest
... The owl:priorVersion property ...
6 - I suggest
... likely not be allowed in a forthcoming release:
6 - Deprecation is part of OWL Lite and is not part of versioning, so I
suggest ending the section
It is important to note that owl:... and owl:... have no
additional semantics and it is up to tool ... intended.
C - I suggest
Appendix C: An Alternative Region Ontology
so that you don't have the connotation, for example, that one is used
on weekdays and the other on weekends.
Received on Monday, 17 March 2003 12:15:55 UTC