DL Query tab problem with overloaded class name

classic Classic list List threaded Threaded
9 messages Options
Reply | Threaded
Open this post in threaded view
|

DL Query tab problem with overloaded class name

boris
Hi,

When a class name is used also as an object property name, the DL query
tab gives a  syntax error when trying to use that class as a query. For
example, if you have a Banana class and a Banana object property, named
the same, and you type in 'Banana' in the DL query tab, it yields a
syntax error.

That's in Protege 4.1.0 Build 209.

Best,
Boris
_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
Reply | Threaded
Open this post in threaded view
|

Re: DL Query tab problem with overloaded class name

Tania Tudorache
I think this is a feature rather than a bug :) It prevents you from
doing something confusing.

Punning can be very tricky, and even if the reasoners support it to some
extent, it can be confusing for the ontology users. In the DL Query tab
case, the class expression parser rejects the expression, and most
likely the expression parser can be fixed to accept it. However, it
would be interesting to see what results you would get back from the
reasoner. I can see the case of punning a class and a property to be
particularly confusing.

I'll discuss with Tim if this should go to the OWL-API list.

Tania

On 06/03/2011 10:38 AM, Iordanov, Borislav (GIC) wrote:

> Hi,
>
> When a class name is used also as an object property name, the DL query
> tab gives a  syntax error when trying to use that class as a query. For
> example, if you have a Banana class and a Banana object property, named
> the same, and you type in 'Banana' in the DL query tab, it yields a
> syntax error.
>
> That's in Protege 4.1.0 Build 209.
>
> Best,
> Boris
> _______________________________________________
> protege-owl mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>
> Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
>

_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
Reply | Threaded
Open this post in threaded view
|

Re: DL Query tab problem with overloaded class name

boris
Hi Tania,

> -----Original Message-----
> From: [hidden email] [mailto:protege-owl-
> [hidden email]] On Behalf Of Tania Tudorache
> Sent: Friday, June 03, 2011 3:36 PM
> To: User support for the Protege-OWL editor
> Subject: Re: [protege-owl] DL Query tab problem with overloaded class
> name
>
> I think this is a feature rather than a bug :) It prevents you from
> doing something confusing.

Confusing to whom? :)

> Punning can be very tricky, and even if the reasoners support it to
> some
> extent, it can be confusing for the ontology users. In the DL Query
tab
> case, the class expression parser rejects the expression, and most
> likely the expression parser can be fixed to accept it. However, it
> would be interesting to see what results you would get back from the
> reasoner. I can see the case of punning a class and a property to be
> particularly confusing.

Interesting! Are there cases where an ambiguity is possible in any of
the OWL and/or SWRL syntaxes? Doesn't syntactic context disambiguate
what kind of entity (class or property or individual) is being referred
to?

Thanks!
Boris

> I'll discuss with Tim if this should go to the OWL-API list.
>
> Tania
>
> On 06/03/2011 10:38 AM, Iordanov, Borislav (GIC) wrote:
> > Hi,
> >
> > When a class name is used also as an object property name, the DL
> query
> > tab gives a  syntax error when trying to use that class as a query.
> For
> > example, if you have a Banana class and a Banana object property,
> named
> > the same, and you type in 'Banana' in the DL query tab, it yields a
> > syntax error.
> >
> > That's in Protege 4.1.0 Build 209.
> >
> > Best,
> > Boris
> > _______________________________________________
> > protege-owl mailing list
> > [hidden email]
> > https://mailman.stanford.edu/mailman/listinfo/protege-owl
> >
> > Instructions for unsubscribing:
> http://protege.stanford.edu/doc/faq.html#01a.03
> >
>
> _______________________________________________
> protege-owl mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>
> Instructions for unsubscribing:
> http://protege.stanford.edu/doc/faq.html#01a.03
_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
Reply | Threaded
Open this post in threaded view
|

Re: DL Query tab problem with overloaded class name

Tania Tudorache
On 06/03/2011 12:52 PM, Iordanov, Borislav (GIC) wrote:

> Hi Tania,
>
>> -----Original Message-----
>> From: [hidden email] [mailto:protege-owl-
>> [hidden email]] On Behalf Of Tania Tudorache
>> Sent: Friday, June 03, 2011 3:36 PM
>> To: User support for the Protege-OWL editor
>> Subject: Re: [protege-owl] DL Query tab problem with overloaded class
>> name
>>
>> I think this is a feature rather than a bug :) It prevents you from
>> doing something confusing.
> Confusing to whom? :)

I think the real question is what you real mean when you use the same
name for a property and a class. What are you trying to model this way?
The reasoner will see them as two separate entities, so you may not get
the inferences you expect. As long as your ontology users are aware of
this, and they understand that the class and property really represent
two separate entities that happen to share a name, then it is fine. But
I can see that many users might expect "more" from two entities sharing
a name. It would be great to have some best practices for modeling with
punning, but I am not aware of any. I know many cases in which a class
and an individual share a name (many times just because the class has a
datatype property assertion), but I have not seen many modeling cases in
which a class and a property are punned (expect by mistake).


>> Punning can be very tricky, and even if the reasoners support it to
>> some
>> extent, it can be confusing for the ontology users. In the DL Query
> tab
>> case, the class expression parser rejects the expression, and most
>> likely the expression parser can be fixed to accept it. However, it
>> would be interesting to see what results you would get back from the
>> reasoner. I can see the case of punning a class and a property to be
>> particularly confusing.
> Interesting! Are there cases where an ambiguity is possible in any of
> the OWL and/or SWRL syntaxes? Doesn't syntactic context disambiguate
> what kind of entity (class or property or individual) is being referred
> to?

I don't know. I would expect that reasoners could use the class
expression grammar to infer if you are talking about the class or a
property, but I am not sure about this, and maybe someone from the
reasoner community could comment on this.

Tania

> Thanks!
> Boris
>
>> I'll discuss with Tim if this should go to the OWL-API list.
>>
>> Tania
>>
>> On 06/03/2011 10:38 AM, Iordanov, Borislav (GIC) wrote:
>>> Hi,
>>>
>>> When a class name is used also as an object property name, the DL
>> query
>>> tab gives a  syntax error when trying to use that class as a query.
>> For
>>> example, if you have a Banana class and a Banana object property,
>> named
>>> the same, and you type in 'Banana' in the DL query tab, it yields a
>>> syntax error.
>>>
>>> That's in Protege 4.1.0 Build 209.
>>>
>>> Best,
>>> Boris
>>> _______________________________________________
>>> protege-owl mailing list
>>> [hidden email]
>>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>>
>>> Instructions for unsubscribing:
>> http://protege.stanford.edu/doc/faq.html#01a.03
>> _______________________________________________
>> protege-owl mailing list
>> [hidden email]
>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>
>> Instructions for unsubscribing:
>> http://protege.stanford.edu/doc/faq.html#01a.03
> _______________________________________________
> protege-owl mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>
> Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
>

_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
Reply | Threaded
Open this post in threaded view
|

Re: DL Query tab problem with overloaded class name

Thomas Russ
In reply to this post by boris

On Jun 3, 2011, at 12:52 PM, Iordanov, Borislav (GIC) wrote:

> Hi Tania,
>
>> -----Original Message-----
>> From: [hidden email] [mailto:protege-owl-
>> [hidden email]] On Behalf Of Tania Tudorache
>> Sent: Friday, June 03, 2011 3:36 PM
>> To: User support for the Protege-OWL editor
>> Subject: Re: [protege-owl] DL Query tab problem with overloaded class
>> name
>>
>> I think this is a feature rather than a bug :) It prevents you from
>> doing something confusing.
>
> Confusing to whom? :)

Users.  Why make the ontology harder to understand, when it would be easy to choose different names for classes and properties.

>
>> Punning can be very tricky, and even if the reasoners support it to
>> some
>> extent, it can be confusing for the ontology users. In the DL Query
> tab
>> case, the class expression parser rejects the expression, and most
>> likely the expression parser can be fixed to accept it. However, it
>> would be interesting to see what results you would get back from the
>> reasoner. I can see the case of punning a class and a property to be
>> particularly confusing.

While I am not fully expert on the detailed semantic considerations, my understanding is that in OWL (and RDF), the point of having a URI is that any given URI refers to exactly one resource or object.  Punning in OWL is used to treat the same object as both an individual and either a class or property.  But the expectation is that the punned items are conceptually the same object.

So, while there are reasons to treat classes or properties as individuals, particularly when doing some type of meta-modeling, it would seem to me that having the same object be both a class and a property would lead to an inconsistency, if for no other reason than classes are unary predicates and properties are binary predicates.  So having an object with multiple arities would be inconsistent.

Now, the OWL 2 DL does allow the use of the same name for properties and classes, but I have to agree with Tania, that I don't see any advantage to doing this, and see potential drawbacks to having to rely on context to disambiguate among the three different entities that you might mean.  At least with the meta-modeling approach, I can see the use cases.  The use case that is identified in the OWL documentation (UML association classes) seems to me a desire to associate the same name with a property and the reification of that property as a class.  But without any of the semantic connections that one would expect to have from such a reification, because there is no semantic connection between the different objects that are being punned.

So, I would strongly recommend not using the same name for classes and properties.


>
> Interesting! Are there cases where an ambiguity is possible in any of
> the OWL and/or SWRL syntaxes? Doesn't syntactic context disambiguate
> what kind of entity (class or property or individual) is being referred
> to?
>
> Thanks!
> Boris
>
>> I'll discuss with Tim if this should go to the OWL-API list.
>>
>> Tania
>>
>> On 06/03/2011 10:38 AM, Iordanov, Borislav (GIC) wrote:
>>> Hi,
>>>
>>> When a class name is used also as an object property name, the DL
>> query
>>> tab gives a  syntax error when trying to use that class as a query.
>> For
>>> example, if you have a Banana class and a Banana object property,
>> named
>>> the same, and you type in 'Banana' in the DL query tab, it yields a
>>> syntax error.
>>>
>>> That's in Protege 4.1.0 Build 209.
>>>
>>> Best,
>>> Boris
>>> _______________________________________________
>>> protege-owl mailing list
>>> [hidden email]
>>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>>
>>> Instructions for unsubscribing:
>> http://protege.stanford.edu/doc/faq.html#01a.03
>>>
>>
>> _______________________________________________
>> protege-owl mailing list
>> [hidden email]
>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>
>> Instructions for unsubscribing:
>> http://protege.stanford.edu/doc/faq.html#01a.03
> _______________________________________________
> protege-owl mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>
> Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03

_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
Reply | Threaded
Open this post in threaded view
|

Re: DL Query tab problem with overloaded class name

boris
In reply to this post by Tania Tudorache
Hi Tania,

Yes, I understand the importance of proper naming and the fact that when
two entities share a name, one expects them to share something more. And
I completely agree with your comments. I just wanted to point out the
fact that the tool (Protege and/or the OWLAPI) is much more confused
than me. First, it doesn't accept something that's legal. Second, the
error message gave me no clue as to the possible cause and it took me a
while to figure it out! It didn't complain about an ambiguous name, but
about syntax.

Thanks!
Boris

> -----Original Message-----
> From: [hidden email] [mailto:protege-owl-
> [hidden email]] On Behalf Of Tania Tudorache
> Sent: Friday, June 03, 2011 5:16 PM
> To: User support for the Protege-OWL editor
> Subject: Re: [protege-owl] DL Query tab problem with overloaded class
> name
>
> On 06/03/2011 12:52 PM, Iordanov, Borislav (GIC) wrote:
> > Hi Tania,
> >
> >> -----Original Message-----
> >> From: [hidden email] [mailto:protege-owl-
> >> [hidden email]] On Behalf Of Tania Tudorache
> >> Sent: Friday, June 03, 2011 3:36 PM
> >> To: User support for the Protege-OWL editor
> >> Subject: Re: [protege-owl] DL Query tab problem with overloaded
> class
> >> name
> >>
> >> I think this is a feature rather than a bug :) It prevents you from
> >> doing something confusing.
> > Confusing to whom? :)
>
> I think the real question is what you real mean when you use the same
> name for a property and a class. What are you trying to model this
way?
> The reasoner will see them as two separate entities, so you may not
get
> the inferences you expect. As long as your ontology users are aware of
> this, and they understand that the class and property really represent
> two separate entities that happen to share a name, then it is fine.
But
> I can see that many users might expect "more" from two entities
sharing
> a name. It would be great to have some best practices for modeling
with
> punning, but I am not aware of any. I know many cases in which a class
> and an individual share a name (many times just because the class has
a

> datatype property assertion), but I have not seen many modeling cases
> in
> which a class and a property are punned (expect by mistake).
>
>
> >> Punning can be very tricky, and even if the reasoners support it to
> >> some
> >> extent, it can be confusing for the ontology users. In the DL Query
> > tab
> >> case, the class expression parser rejects the expression, and most
> >> likely the expression parser can be fixed to accept it. However, it
> >> would be interesting to see what results you would get back from
the
> >> reasoner. I can see the case of punning a class and a property to
be
> >> particularly confusing.
> > Interesting! Are there cases where an ambiguity is possible in any
of

> > the OWL and/or SWRL syntaxes? Doesn't syntactic context disambiguate
> > what kind of entity (class or property or individual) is being
> referred
> > to?
>
> I don't know. I would expect that reasoners could use the class
> expression grammar to infer if you are talking about the class or a
> property, but I am not sure about this, and maybe someone from the
> reasoner community could comment on this.
>
> Tania
>
> > Thanks!
> > Boris
> >
> >> I'll discuss with Tim if this should go to the OWL-API list.
> >>
> >> Tania
> >>
> >> On 06/03/2011 10:38 AM, Iordanov, Borislav (GIC) wrote:
> >>> Hi,
> >>>
> >>> When a class name is used also as an object property name, the DL
> >> query
> >>> tab gives a  syntax error when trying to use that class as a
query.
> >> For
> >>> example, if you have a Banana class and a Banana object property,
> >> named
> >>> the same, and you type in 'Banana' in the DL query tab, it yields
a

> >>> syntax error.
> >>>
> >>> That's in Protege 4.1.0 Build 209.
> >>>
> >>> Best,
> >>> Boris
> >>> _______________________________________________
> >>> protege-owl mailing list
> >>> [hidden email]
> >>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
> >>>
> >>> Instructions for unsubscribing:
> >> http://protege.stanford.edu/doc/faq.html#01a.03
> >> _______________________________________________
> >> protege-owl mailing list
> >> [hidden email]
> >> https://mailman.stanford.edu/mailman/listinfo/protege-owl
> >>
> >> Instructions for unsubscribing:
> >> http://protege.stanford.edu/doc/faq.html#01a.03
> > _______________________________________________
> > protege-owl mailing list
> > [hidden email]
> > https://mailman.stanford.edu/mailman/listinfo/protege-owl
> >
> > Instructions for unsubscribing:
> http://protege.stanford.edu/doc/faq.html#01a.03
> >
>
> _______________________________________________
> protege-owl mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>
> Instructions for unsubscribing:
> http://protege.stanford.edu/doc/faq.html#01a.03
_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
Reply | Threaded
Open this post in threaded view
|

Re: DL Query tab problem with overloaded class name

Tania Tudorache
I know, and I agree with you. The parser should have worked in this
case, and if not, it should have given you better feedback of what it is
wrong.

The class parser is part of the OWL-API. I will discuss with Tim, when
he gets back, whether to submit a bug report to the OWL-API mailing list.

Tania

On 06/05/2011 11:33 AM, Iordanov, Borislav (GIC) wrote:

> Hi Tania,
>
> Yes, I understand the importance of proper naming and the fact that when
> two entities share a name, one expects them to share something more. And
> I completely agree with your comments. I just wanted to point out the
> fact that the tool (Protege and/or the OWLAPI) is much more confused
> than me. First, it doesn't accept something that's legal. Second, the
> error message gave me no clue as to the possible cause and it took me a
> while to figure it out! It didn't complain about an ambiguous name, but
> about syntax.
>
> Thanks!
> Boris
>
>> -----Original Message-----
>> From: [hidden email] [mailto:protege-owl-
>> [hidden email]] On Behalf Of Tania Tudorache
>> Sent: Friday, June 03, 2011 5:16 PM
>> To: User support for the Protege-OWL editor
>> Subject: Re: [protege-owl] DL Query tab problem with overloaded class
>> name
>>
>> On 06/03/2011 12:52 PM, Iordanov, Borislav (GIC) wrote:
>>> Hi Tania,
>>>
>>>> -----Original Message-----
>>>> From: [hidden email] [mailto:protege-owl-
>>>> [hidden email]] On Behalf Of Tania Tudorache
>>>> Sent: Friday, June 03, 2011 3:36 PM
>>>> To: User support for the Protege-OWL editor
>>>> Subject: Re: [protege-owl] DL Query tab problem with overloaded
>> class
>>>> name
>>>>
>>>> I think this is a feature rather than a bug :) It prevents you from
>>>> doing something confusing.
>>> Confusing to whom? :)
>> I think the real question is what you real mean when you use the same
>> name for a property and a class. What are you trying to model this
> way?
>> The reasoner will see them as two separate entities, so you may not
> get
>> the inferences you expect. As long as your ontology users are aware of
>> this, and they understand that the class and property really represent
>> two separate entities that happen to share a name, then it is fine.
> But
>> I can see that many users might expect "more" from two entities
> sharing
>> a name. It would be great to have some best practices for modeling
> with
>> punning, but I am not aware of any. I know many cases in which a class
>> and an individual share a name (many times just because the class has
> a
>> datatype property assertion), but I have not seen many modeling cases
>> in
>> which a class and a property are punned (expect by mistake).
>>
>>
>>>> Punning can be very tricky, and even if the reasoners support it to
>>>> some
>>>> extent, it can be confusing for the ontology users. In the DL Query
>>> tab
>>>> case, the class expression parser rejects the expression, and most
>>>> likely the expression parser can be fixed to accept it. However, it
>>>> would be interesting to see what results you would get back from
> the
>>>> reasoner. I can see the case of punning a class and a property to
> be
>>>> particularly confusing.
>>> Interesting! Are there cases where an ambiguity is possible in any
> of
>>> the OWL and/or SWRL syntaxes? Doesn't syntactic context disambiguate
>>> what kind of entity (class or property or individual) is being
>> referred
>>> to?
>> I don't know. I would expect that reasoners could use the class
>> expression grammar to infer if you are talking about the class or a
>> property, but I am not sure about this, and maybe someone from the
>> reasoner community could comment on this.
>>
>> Tania
>>
>>> Thanks!
>>> Boris
>>>
>>>> I'll discuss with Tim if this should go to the OWL-API list.
>>>>
>>>> Tania
>>>>
>>>> On 06/03/2011 10:38 AM, Iordanov, Borislav (GIC) wrote:
>>>>> Hi,
>>>>>
>>>>> When a class name is used also as an object property name, the DL
>>>> query
>>>>> tab gives a  syntax error when trying to use that class as a
> query.
>>>> For
>>>>> example, if you have a Banana class and a Banana object property,
>>>> named
>>>>> the same, and you type in 'Banana' in the DL query tab, it yields
> a
>>>>> syntax error.
>>>>>
>>>>> That's in Protege 4.1.0 Build 209.
>>>>>
>>>>> Best,
>>>>> Boris
>>>>> _______________________________________________
>>>>> protege-owl mailing list
>>>>> [hidden email]
>>>>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>>>>
>>>>> Instructions for unsubscribing:
>>>> http://protege.stanford.edu/doc/faq.html#01a.03
>>>> _______________________________________________
>>>> protege-owl mailing list
>>>> [hidden email]
>>>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>>>
>>>> Instructions for unsubscribing:
>>>> http://protege.stanford.edu/doc/faq.html#01a.03
>>> _______________________________________________
>>> protege-owl mailing list
>>> [hidden email]
>>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>>
>>> Instructions for unsubscribing:
>> http://protege.stanford.edu/doc/faq.html#01a.03
>> _______________________________________________
>> protege-owl mailing list
>> [hidden email]
>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>
>> Instructions for unsubscribing:
>> http://protege.stanford.edu/doc/faq.html#01a.03
> _______________________________________________
> protege-owl mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>
> Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
>

_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
Reply | Threaded
Open this post in threaded view
|

Re: DL Query tab problem with overloaded class name

Tania Tudorache
Follow up on this issue. There is already a bug submitted to the OWL-API
tracker about the Manchester syntax parser not handling well classes
when they are also punned as object properties:

http://sourceforge.net/tracker/index.php?func=detail&aid=2947144&group_id=90989&atid=595534

Tania

On 06/06/2011 03:54 PM, Tania Tudorache wrote:

> I know, and I agree with you. The parser should have worked in this
> case, and if not, it should have given you better feedback of what it
> is wrong.
>
> The class parser is part of the OWL-API. I will discuss with Tim, when
> he gets back, whether to submit a bug report to the OWL-API mailing list.
>
> Tania
>
> On 06/05/2011 11:33 AM, Iordanov, Borislav (GIC) wrote:
>> Hi Tania,
>>
>> Yes, I understand the importance of proper naming and the fact that when
>> two entities share a name, one expects them to share something more. And
>> I completely agree with your comments. I just wanted to point out the
>> fact that the tool (Protege and/or the OWLAPI) is much more confused
>> than me. First, it doesn't accept something that's legal. Second, the
>> error message gave me no clue as to the possible cause and it took me a
>> while to figure it out! It didn't complain about an ambiguous name, but
>> about syntax.
>>
>> Thanks!
>> Boris
>>
>>> -----Original Message-----
>>> From: [hidden email] [mailto:protege-owl-
>>> [hidden email]] On Behalf Of Tania Tudorache
>>> Sent: Friday, June 03, 2011 5:16 PM
>>> To: User support for the Protege-OWL editor
>>> Subject: Re: [protege-owl] DL Query tab problem with overloaded class
>>> name
>>>
>>> On 06/03/2011 12:52 PM, Iordanov, Borislav (GIC) wrote:
>>>> Hi Tania,
>>>>
>>>>> -----Original Message-----
>>>>> From: [hidden email] [mailto:protege-owl-
>>>>> [hidden email]] On Behalf Of Tania Tudorache
>>>>> Sent: Friday, June 03, 2011 3:36 PM
>>>>> To: User support for the Protege-OWL editor
>>>>> Subject: Re: [protege-owl] DL Query tab problem with overloaded
>>> class
>>>>> name
>>>>>
>>>>> I think this is a feature rather than a bug :) It prevents you from
>>>>> doing something confusing.
>>>> Confusing to whom? :)
>>> I think the real question is what you real mean when you use the same
>>> name for a property and a class. What are you trying to model this
>> way?
>>> The reasoner will see them as two separate entities, so you may not
>> get
>>> the inferences you expect. As long as your ontology users are aware of
>>> this, and they understand that the class and property really represent
>>> two separate entities that happen to share a name, then it is fine.
>> But
>>> I can see that many users might expect "more" from two entities
>> sharing
>>> a name. It would be great to have some best practices for modeling
>> with
>>> punning, but I am not aware of any. I know many cases in which a class
>>> and an individual share a name (many times just because the class has
>> a
>>> datatype property assertion), but I have not seen many modeling cases
>>> in
>>> which a class and a property are punned (expect by mistake).
>>>
>>>
>>>>> Punning can be very tricky, and even if the reasoners support it to
>>>>> some
>>>>> extent, it can be confusing for the ontology users. In the DL Query
>>>> tab
>>>>> case, the class expression parser rejects the expression, and most
>>>>> likely the expression parser can be fixed to accept it. However, it
>>>>> would be interesting to see what results you would get back from
>> the
>>>>> reasoner. I can see the case of punning a class and a property to
>> be
>>>>> particularly confusing.
>>>> Interesting! Are there cases where an ambiguity is possible in any
>> of
>>>> the OWL and/or SWRL syntaxes? Doesn't syntactic context disambiguate
>>>> what kind of entity (class or property or individual) is being
>>> referred
>>>> to?
>>> I don't know. I would expect that reasoners could use the class
>>> expression grammar to infer if you are talking about the class or a
>>> property, but I am not sure about this, and maybe someone from the
>>> reasoner community could comment on this.
>>>
>>> Tania
>>>
>>>> Thanks!
>>>> Boris
>>>>
>>>>> I'll discuss with Tim if this should go to the OWL-API list.
>>>>>
>>>>> Tania
>>>>>
>>>>> On 06/03/2011 10:38 AM, Iordanov, Borislav (GIC) wrote:
>>>>>> Hi,
>>>>>>
>>>>>> When a class name is used also as an object property name, the DL
>>>>> query
>>>>>> tab gives a  syntax error when trying to use that class as a
>> query.
>>>>> For
>>>>>> example, if you have a Banana class and a Banana object property,
>>>>> named
>>>>>> the same, and you type in 'Banana' in the DL query tab, it yields
>> a
>>>>>> syntax error.
>>>>>>
>>>>>> That's in Protege 4.1.0 Build 209.
>>>>>>
>>>>>> Best,
>>>>>> Boris
>>>>>> _______________________________________________
>>>>>> protege-owl mailing list
>>>>>> [hidden email]
>>>>>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>>>>>
>>>>>> Instructions for unsubscribing:
>>>>> http://protege.stanford.edu/doc/faq.html#01a.03
>>>>> _______________________________________________
>>>>> protege-owl mailing list
>>>>> [hidden email]
>>>>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>>>>
>>>>> Instructions for unsubscribing:
>>>>> http://protege.stanford.edu/doc/faq.html#01a.03
>>>> _______________________________________________
>>>> protege-owl mailing list
>>>> [hidden email]
>>>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>>>
>>>> Instructions for unsubscribing:
>>> http://protege.stanford.edu/doc/faq.html#01a.03
>>> _______________________________________________
>>> protege-owl mailing list
>>> [hidden email]
>>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>>
>>> Instructions for unsubscribing:
>>> http://protege.stanford.edu/doc/faq.html#01a.03
>> _______________________________________________
>> protege-owl mailing list
>> [hidden email]
>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>>
>> Instructions for unsubscribing:
>> http://protege.stanford.edu/doc/faq.html#01a.03
>>
>
> _______________________________________________
> protege-owl mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>
> Instructions for unsubscribing:
> http://protege.stanford.edu/doc/faq.html#01a.03
>

_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03
Reply | Threaded
Open this post in threaded view
|

Re: DL Query tab problem with overloaded class name

boris
Thanks, Tania, that's good to know.

Cheers,
Boris

> -----Original Message-----
> From: [hidden email] [mailto:protege-owl-
> [hidden email]] On Behalf Of Tania Tudorache
> Sent: Thursday, June 09, 2011 2:00 PM
> To: User support for the Protege-OWL editor
> Subject: Re: [protege-owl] DL Query tab problem with overloaded class
> name
>
> Follow up on this issue. There is already a bug submitted to the OWL-
> API
> tracker about the Manchester syntax parser not handling well classes
> when they are also punned as object properties:
>
>
http://sourceforge.net/tracker/index.php?func=detail&aid=2947144&group_
> id=90989&atid=595534
>
> Tania
>
> On 06/06/2011 03:54 PM, Tania Tudorache wrote:
> > I know, and I agree with you. The parser should have worked in this
> > case, and if not, it should have given you better feedback of what
it

> > is wrong.
> >
> > The class parser is part of the OWL-API. I will discuss with Tim,
> when
> > he gets back, whether to submit a bug report to the OWL-API mailing
> list.
> >
> > Tania
> >
> > On 06/05/2011 11:33 AM, Iordanov, Borislav (GIC) wrote:
> >> Hi Tania,
> >>
> >> Yes, I understand the importance of proper naming and the fact that
> when
> >> two entities share a name, one expects them to share something
more.
> And
> >> I completely agree with your comments. I just wanted to point out
> the
> >> fact that the tool (Protege and/or the OWLAPI) is much more
confused

> >> than me. First, it doesn't accept something that's legal. Second,
> the
> >> error message gave me no clue as to the possible cause and it took
> me a
> >> while to figure it out! It didn't complain about an ambiguous name,
> but
> >> about syntax.
> >>
> >> Thanks!
> >> Boris
> >>
> >>> -----Original Message-----
> >>> From: [hidden email] [mailto:protege-owl-
> >>> [hidden email]] On Behalf Of Tania Tudorache
> >>> Sent: Friday, June 03, 2011 5:16 PM
> >>> To: User support for the Protege-OWL editor
> >>> Subject: Re: [protege-owl] DL Query tab problem with overloaded
> class
> >>> name
> >>>
> >>> On 06/03/2011 12:52 PM, Iordanov, Borislav (GIC) wrote:
> >>>> Hi Tania,
> >>>>
> >>>>> -----Original Message-----
> >>>>> From: [hidden email]
[mailto:protege-owl-

> >>>>> [hidden email]] On Behalf Of Tania Tudorache
> >>>>> Sent: Friday, June 03, 2011 3:36 PM
> >>>>> To: User support for the Protege-OWL editor
> >>>>> Subject: Re: [protege-owl] DL Query tab problem with overloaded
> >>> class
> >>>>> name
> >>>>>
> >>>>> I think this is a feature rather than a bug :) It prevents you
> from
> >>>>> doing something confusing.
> >>>> Confusing to whom? :)
> >>> I think the real question is what you real mean when you use the
> same
> >>> name for a property and a class. What are you trying to model this
> >> way?
> >>> The reasoner will see them as two separate entities, so you may
not
> >> get
> >>> the inferences you expect. As long as your ontology users are
aware
> of
> >>> this, and they understand that the class and property really
> represent
> >>> two separate entities that happen to share a name, then it is
fine.

> >> But
> >>> I can see that many users might expect "more" from two entities
> >> sharing
> >>> a name. It would be great to have some best practices for modeling
> >> with
> >>> punning, but I am not aware of any. I know many cases in which a
> class
> >>> and an individual share a name (many times just because the class
> has
> >> a
> >>> datatype property assertion), but I have not seen many modeling
> cases
> >>> in
> >>> which a class and a property are punned (expect by mistake).
> >>>
> >>>
> >>>>> Punning can be very tricky, and even if the reasoners support it
> to
> >>>>> some
> >>>>> extent, it can be confusing for the ontology users. In the DL
> Query
> >>>> tab
> >>>>> case, the class expression parser rejects the expression, and
> most
> >>>>> likely the expression parser can be fixed to accept it. However,
> it
> >>>>> would be interesting to see what results you would get back from
> >> the
> >>>>> reasoner. I can see the case of punning a class and a property
to
> >> be
> >>>>> particularly confusing.
> >>>> Interesting! Are there cases where an ambiguity is possible in
any
> >> of
> >>>> the OWL and/or SWRL syntaxes? Doesn't syntactic context
> disambiguate
> >>>> what kind of entity (class or property or individual) is being
> >>> referred
> >>>> to?
> >>> I don't know. I would expect that reasoners could use the class
> >>> expression grammar to infer if you are talking about the class or
a

> >>> property, but I am not sure about this, and maybe someone from the
> >>> reasoner community could comment on this.
> >>>
> >>> Tania
> >>>
> >>>> Thanks!
> >>>> Boris
> >>>>
> >>>>> I'll discuss with Tim if this should go to the OWL-API list.
> >>>>>
> >>>>> Tania
> >>>>>
> >>>>> On 06/03/2011 10:38 AM, Iordanov, Borislav (GIC) wrote:
> >>>>>> Hi,
> >>>>>>
> >>>>>> When a class name is used also as an object property name, the
> DL
> >>>>> query
> >>>>>> tab gives a  syntax error when trying to use that class as a
> >> query.
> >>>>> For
> >>>>>> example, if you have a Banana class and a Banana object
> property,
> >>>>> named
> >>>>>> the same, and you type in 'Banana' in the DL query tab, it
> yields
> >> a
> >>>>>> syntax error.
> >>>>>>
> >>>>>> That's in Protege 4.1.0 Build 209.
> >>>>>>
> >>>>>> Best,
> >>>>>> Boris
> >>>>>> _______________________________________________
> >>>>>> protege-owl mailing list
> >>>>>> [hidden email]
> >>>>>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
> >>>>>>
> >>>>>> Instructions for unsubscribing:
> >>>>> http://protege.stanford.edu/doc/faq.html#01a.03
> >>>>> _______________________________________________
> >>>>> protege-owl mailing list
> >>>>> [hidden email]
> >>>>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
> >>>>>
> >>>>> Instructions for unsubscribing:
> >>>>> http://protege.stanford.edu/doc/faq.html#01a.03
> >>>> _______________________________________________
> >>>> protege-owl mailing list
> >>>> [hidden email]
> >>>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
> >>>>
> >>>> Instructions for unsubscribing:
> >>> http://protege.stanford.edu/doc/faq.html#01a.03
> >>> _______________________________________________
> >>> protege-owl mailing list
> >>> [hidden email]
> >>> https://mailman.stanford.edu/mailman/listinfo/protege-owl
> >>>
> >>> Instructions for unsubscribing:
> >>> http://protege.stanford.edu/doc/faq.html#01a.03
> >> _______________________________________________
> >> protege-owl mailing list
> >> [hidden email]
> >> https://mailman.stanford.edu/mailman/listinfo/protege-owl
> >>
> >> Instructions for unsubscribing:
> >> http://protege.stanford.edu/doc/faq.html#01a.03
> >>
> >
> > _______________________________________________
> > protege-owl mailing list
> > [hidden email]
> > https://mailman.stanford.edu/mailman/listinfo/protege-owl
> >
> > Instructions for unsubscribing:
> > http://protege.stanford.edu/doc/faq.html#01a.03
> >
>
> _______________________________________________
> protege-owl mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-owl
>
> Instructions for unsubscribing:
> http://protege.stanford.edu/doc/faq.html#01a.03
_______________________________________________
protege-owl mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03