Re: proposal to close (as RESOLVED) ISSUE-90 (class and property deprecation)

Hi Jim,

You're not entirely spot-on about the numbers -- owl:DeprecatedClass  
has only about 400 hits in google (mainly documentation), and  
Swoogle's hits are primarily RDF definitions of the OWL vocabulary --  
but I don't think numbers are particularly relevant anyway, it is  
rather whether the people that *do* use them, use them in some  
important way, and are particularly concerned about their use in  
combination with DL semantics...

Do you have any ideas on how we can get a better overview of usage?

-Rinke



On 24 jan 2008, at 15:53, Jim Hendler wrote:

> Alan - with due respect, I do not believe I have any obligation to  
> "bolster support" for not deleting features - see my actual argument  
> in the email --
> I'm very troubled that this WG somehow thinks we know every  
> implementation and project in OWL and what is in it?  How arrogant!   
> What rules are we playing by?  I remind me there are a lot more  
> implementations of OWL out there than the few complete reasoners  
> people know about, that there are OWL projects behind firewalls,  
> there are OWL projects you won't know about with an NDA, there are  
> projects interacting with OWL in the enterprise.
> Let me suggest that if an open web Google search for  
> "DeprecatedClass" finds 2,340 hits (as of this morning), and that  
> doesn't include anything behind firewalls and such - maybe there are  
> some users.
> OK. now that I've flamed and pointed out that even if I couldn't  
> identify any we would still be duty bound not to change -- I  
> actually do indeed know some users of this - in at least one project  
> I know, a software development group in a company I worked with is  
> using OWL to track software - they created subclasses of  
> deprecatedClass that they use to represent software from version to  
> version - this allows them to issue warnings from a code analyzer  
> when a user has use of a piece of code (they have classes  
> representing various aspects of code).
> Swoogle also finds about 180 uses, most of them from a project at  
> one of the Netherlands
> I also know several DoD projects that are using OWL and do use this,  
> as it was intended, for ontology developers to let other ontology  
> developers know that a particular class or property is intended not  
> to be used.  I believe they are (or were when I heard) using Protege- 
> owl which supported these -- Oh look, here's one - top hit for  
> "deprecatedClass Protege" - as of a year ago the National Cancer  
> Institute Thesaurus (which I believe several of the people on this  
> list claim to track) was using it -- https://mailman.stanford.edu/pipermail/protege-owl/2007-January/001080.html
> Hmm, at this point I'm finding them using just Google - didn't  
> anyone else feel compelled to look before asserting that there were  
> no users?
> -JH
>
>
> On Jan 23, 2008, at 11:49 PM, Alan Ruttenberg wrote:
>
>>
>> On Jan 23, 2008, at 10:42 PM, Jim Hendler wrote:
>>
>>> So for what it is worth, as usual, I have exactly the opposite  
>>> opinion on this as Peter - I think we should close this by leaving  
>>> deprecateion as it is -- yes it is little used, but we did have  
>>> support from it from some developers in OWL 1.0
>>
>> Hi Jim,
>>
>> It would be of interest to be able to cite actual support and  
>> usage. In the meeting none present could remember seeing any, which  
>> is why we suggested that Peter write up the proposal (we suggested  
>> his option 1 so as to not add insult to injury by having it be that  
>> someone who happened to have the non-semantic use of deprecation  
>> wouldn't be dunned with a push into OWL Full solely for this  
>> offense).
>>
>> But if you could dig up some evidence of actual use, it would  
>> bolster the case that there is indeed someone who would be affected  
>> by this.
>>
>> Regards,
>>
>> -Alan
>>
>>> , it has no semantic impact (and should continue to have none) --  
>>> basically, it is a human-readable way of indicating the intent for  
>>> new versions to  overwrite old.  It does no harm that I can find.   
>>> The charter makes it clear that "Backwards compatibility with OWL  
>>> is of great importance" and mandates that we don't add new  
>>> features that break compatibility if there is any doubt of the  
>>> need, I'd suggest that this implies we should also not remove any  
>>> old features unless we can show real need to do so.
>>>  So I propose we close Issue-90 as resolved by saying that no  
>>> change is made from OWL 1.0 to OWL 1.1 to owl:DeprecatedClass and  
>>> owl:DeprecatedProperty.
>>> Syntax: no change
>>> Semantics: no change
>>> RDF mapping: no change
>>> backward compatibility: maintained
>>>  -JH
>>>
>>>
>>> On Jan 23, 2008, at 2:20 PM, Peter F. Patel-Schneider wrote:
>>>
>>>>
>>>> As I mentioned in the teleconference on 23 January 2008,  I  
>>>> propose to
>>>> close ISSUE-90 by deprecating deprecation.
>>>>
>>>> This requires the following changes:
>>>>
>>>> Syntax: Add a note to the Differences section saying that  
>>>> deprecation of
>>>> 	classes, datatypes, and properties is deprecated and is not a
>>>> 	part of the functional syntax or structural specification.
>>>>
>>>> 	No other change.
>>>>
>>>> Semantics: No change.
>>>>
>>>> RDF Mapping: Add a new section at the (that will be much expanded  
>>>> later,
>>>>    	     probably) to mention that owl:DeprecatedClass and
>>>>    	     owl:DeprecatedProperty are not part of OWL 1.1.
>>>>
>>>>   OPTION 1: Add a paragraph to Section 3 saying that triples of the
>>>> 	     form x rdf:type owl:DeprecatedClass where Type(x) contains
>>>> 	     owl:Class or rdfs:Datatype, or of the form x rdf:type
>>>> 	     owl:DeprecatedProperty where Type(x) contains
>>>> 	     owl:ObjectProperty or owl:DatatypeProperty or
>>>> 	     owl:AnnotationProperty are removed
>>>>
>>>>   OPTION 2: No change to Section 3, which means that use of
>>>>   	     owl:DeprecatedClass or owl:DeprecatedProperty is not in OWL
>>>>   	     1.1.
>>>>
>>>> I much prefer OPTION 2.
>>>>
>>>>
>>>> peter
>>>>
>>>
>>> "If we knew what we were doing, it wouldn't be called research,  
>>> would it?." - Albert Einstein
>>>
>>> Prof James Hendler				http://www.cs.rpi.edu/~hendler
>>> Tetherless World Constellation Chair
>>> Computer Science Dept
>>> Rensselaer Polytechnic Institute, Troy NY 12180
>>>
>>>
>>>
>>>
>>
>
> "If we knew what we were doing, it wouldn't be called research,  
> would it?." - Albert Einstein
>
> Prof James Hendler				http://www.cs.rpi.edu/~hendler
> Tetherless World Constellation Chair
> Computer Science Dept
> Rensselaer Polytechnic Institute, Troy NY 12180
>
>
>
>

-----------------------------------------------
Drs. Rinke Hoekstra

Email: hoekstra@uva.nl    Skype:  rinkehoekstra
Phone: +31-20-5253499     Fax:   +31-20-5253495
Web:   http://www.leibnizcenter.org/users/rinke

Leibniz Center for Law,          Faculty of Law
University of Amsterdam,            PO Box 1030
1000 BA  Amsterdam,             The Netherlands
-----------------------------------------------

Received on Thursday, 24 January 2008 15:33:57 UTC