Warning:
This wiki has been archived and is now read-only.
LC Responses/SS1a
See also FH6
| Personally, I feel OWL2 is a wrap-up of different recent efforts in the
OWL community from different research forces, such as EL from Dresden, QL and RL from Manchester/Oxford. Given the usage of OWL 1.0 is quite limited in the industry compare to the usage of RDF, it may cost many extra efforts and is very challenging to teach system developers to use new OWL2, in particular, identifying different subsets of OWL2 for developers with limited logic background. Nevertheless, new features on DataProperty related predicates could be useful for semantic application developers in defining and reasoning over their data and metadata. |
PeterPatel-Schneider 02:21, 14 February 2009 (UTC)
[[Draft Response: This response is for both parts of LC comment 26.]
Dear Susie,
Thank you for your message
http://lists.w3.org/Archives/Public/public-owl-comments/2009Jan/0033.html
on the OWL 2 Web Ontology Language last call drafts.
1/ Concerning OWL 2 Profiles:
The OWL 2 effort was designed to extend the constructs of OWL to encompass those that users had asked for, that researchers could specify, and that implementers had or could implement. It is thus definitely the case that OWL 2 is supported by research. However, the driving force was much more users and implementers.
The OWL profiles had a similar genesis. If the driving force behind the OWL profiles had primarily been research, then there could have been many more profiles, and the profiles would have had a different focus.
This is particularly the case for OWL RL. OWL RL is designed to capture the essence of several partial implementations of OWL functionality by means of forward chaining rules. Previously all that could be said about these implementations was that they were partial implementations of OWL. OWL RL provides a much more complete characterization for rule-based implementations of OWL. Yes, there are formal results underlying OWL RL, but these formal results are descriptive of the extant implementations instead of being driving forces for the design of OWL RL.
OWL EL and OWL QL also do have a formal basis. However they again are attempts to capture existing implementation techniques and existing ontologies.
The WG also expects that there will be tools that analyze OWL 2 ontologies and determine which profile they fit into, just as was the case in the past.
In any case, the OWL 2 profiles are simply there for those who may be interested in exploiting desirable characteristics of implementations of the profiles. If one does not care about these, there is no need to consider the profiles at all.
The OWL WG does not intend to make any changes in response to this part of your message.
2/ Concerning the Structural Specification and Functional-Style Syntax:
Thank you for your kind words on the introduction and other parts of this document. Also thank you for bringing to our attention the extraneous word in Section 2.3.
There is no notion of ownership of entities in OWL, so "ownership" is not transferred with imports. It is the case, however, that entities in imported ontologies can be used in the importing ontology just as if the contents of the imported ontology were part of the importing ontology.
The situation with -0 and +0 is unique, and is dictated by the treatment of floats in XML Schema datatypes. There are no other similar situations. The literals for booleans are similarly dictated by XML Schema datatypes. If "yes" and "no" are added to the XML Schema datatype boolean, they will then be available in OWL.
3/ Mapping to RDF Graphs:
It is definitely the case that mapping from the Functional Syntax to an RDF graph and back again does not affect the meaning of an OWL 2 ontology. The OWL WG expended considerable effort to make this mapping as general as possible and to specify it in more detail than previously.
Please acknowledge receipt of this email to
<mailto:public-owl-comments@w3.org> (replying to this email should
suffice). In your acknowledgment please let us know whether or not you
are satisfied with the working group's response to your comment.
Regards, Peter F. Patel-Schneider on behalf of the W3C OWL Working Group
PeterPatel-Schneider 16:18, 19 February 2009 (UTC)
[Response for LC Comment 26:]
Dear Susie,
Thank you for your message
http://lists.w3.org/Archives/Public/public-owl-comments/2009Jan/0033.html
on the OWL 2 Web Ontology Language last call drafts. part of which was supported by Frank van Harmelen in
http://lists.w3.org/Archives/Public/public-owl-comments/2009Jan/0042.html
1/ Concerning OWL 2 Profiles:
The OWL 2 effort was designed to extend the constructs of OWL to encompass those that users had asked for, that researchers could specify, and that implementers had or could implement. It is thus definitely the case that OWL 2 is supported by research. However, the driving force was much more users and implementers.
The OWL profiles had a similar genesis. If the driving force behind the OWL profiles had primarily been research, then there could have been many more profiles, and the profiles would have had a different focus.
This is particularly the case for OWL RL. OWL RL is designed to capture the essence of several partial implementations of OWL functionality by means of forward chaining rules. Previously all that could be said about these implementations was that they were partial implementations of OWL. OWL RL provides a much more complete characterization for rule-based implementations of OWL. Yes, there are formal results underlying OWL RL, but these formal results are descriptive of the extant implementations instead of being driving forces for the design of OWL RL.
OWL EL and OWL QL also do have a formal basis. However they again are attempts to capture existing implementation techniques and existing ontologies.
The WG also expects that there will be tools that analyze OWL 2 ontologies and determine which profile they fit into, just as was the case in the past.
In any case, the OWL 2 profiles are simply there for those who may be interested taking exploiting desirable characteristics of implementations of the profiles. If one does not care about these, there is no need to consider the profiles at all. The Primer document discusses the different profiles, and provides some guidance as to which profile to use.
2/ Concerning the Structural Specification and Functional-Style Syntax:
Thank you for your kind words on the introduction and other parts of this document. Also thank you for bringing to our attention the extraneous word in Section 2.3.
There is no notion of ownership of entities in OWL, so "ownership" is not transferred with imports. It is the case, however, that entities in imported ontologies can be used in the importing ontology just as if the contents of the imported ontology were part of the importing ontology.
The situation with -0 and +0 is unique, and is dictated by the treatment of floats in XML Schema datatypes. There are no other similar situations. The literals for booleans are similarly dictated by XML Schema datatypes. If "yes" and "no" are added to the XML Schema datatype boolean, they will then be available in OWL.
3/ Mapping to RDF Graphs:
It is definitely the case that mapping from the Functional Syntax to an RDF graph and back again does not affect the meaning of an OWL 2 ontology. The OWL WG expended considerable effort to make this mapping as general as possible and to specify it in more detail than previously.
The OWL WG does not intend to make any changes in response to your
message, except that the Primer is still not in Last Call and its
discussion on profiles may be improved in response to your concerns.
Please acknowledge receipt of this email to
<mailto:public-owl-comments@w3.org> (replying to this email should
suffice). In your acknowledgment please let us know whether or not you
are satisfied with the working group's response to your comment.
Regards, Peter F. Patel-Schneider on behalf of the W3C OWL Working Group