Music Ontology Issues
93 views
Skip to first unread message
Carlos Bobed
Jul 27, 2012, 9:52:13 AM7/27/12
to music-ontology-sp...@googlegroups.com
Hello everyone!
I was studying the Music Ontology to use the vocabulary it provides
and I've found out that the current version:
http://motools.sourceforge.net/doc/musicontology.rdfs
as it is in the Web cannot be loaded directly by editors such as
Protegé or NeOnToolkit. There is a problem with importing the
similarity ontology (MuSim): there is an ontology that imports it and
references its namespace http://purl.org/ontology/similarity/ instead
of the physical URL. I've found a workaround in the editors I use,
this is, to load firstly by hand the similarity ontology in the
workspace and then load the Music Ontology in the same workspace (this
way, when it imports the namespace, it takes it from the loaded
version).
Besides, a question has come into my mind once I've managed to load
it: is this ontology oriented to perform any reasoning task or just to
specify a vocabulary for publishing Linked Data? I mean, once loaded,
I've classified it with three different reasoners (Pellet, HermiT and
Fact++) and all of them tell me that Association (the class from
similarity ontology) is equivalent to Thing, and therefore, everything
in the ontology has to be considered an association. If I keep the
expressivity of my system just at RDF level, it might not be
important; however, if I want to perform some reasoning over a
classified A-Box, this might be quite harmful. What kind of tasks were
thought of to be performed with this ontology?
Thank you very much in advance,
Carlos Bobed Lisbona
University of Zaragoza, Spain
and I've found out that the current version:
http://motools.sourceforge.net/doc/musicontology.rdfs
as it is in the Web cannot be loaded directly by editors such as
Protegé or NeOnToolkit. There is a problem with importing the
similarity ontology (MuSim): there is an ontology that imports it and
references its namespace http://purl.org/ontology/similarity/ instead
of the physical URL. I've found a workaround in the editors I use,
this is, to load firstly by hand the similarity ontology in the
workspace and then load the Music Ontology in the same workspace (this
way, when it imports the namespace, it takes it from the loaded
version).
Besides, a question has come into my mind once I've managed to load
it: is this ontology oriented to perform any reasoning task or just to
specify a vocabulary for publishing Linked Data? I mean, once loaded,
I've classified it with three different reasoners (Pellet, HermiT and
Fact++) and all of them tell me that Association (the class from
similarity ontology) is equivalent to Thing, and therefore, everything
in the ontology has to be considered an association. If I keep the
expressivity of my system just at RDF level, it might not be
important; however, if I want to perform some reasoning over a
classified A-Box, this might be quite harmful. What kind of tasks were
thought of to be performed with this ontology?
Thank you very much in advance,
Carlos Bobed Lisbona
University of Zaragoza, Spain
Bob Ferris
Jul 27, 2012, 10:15:18 AM7/27/12
to music-ontology-sp...@googlegroups.com
Hi Carlos,
thanks a lot for you interest in the Music Ontology.
Here is only a short answer from me for now:
(see my inline comments)
On 07/27/2012 03:52 PM, Carlos Bobed wrote:
> Hello everyone!
>
> I was studying the Music Ontology to use the vocabulary it provides
> and I've found out that the current version:
>
> http://motools.sourceforge.net/doc/musicontology.rdfs
>
> as it is in the Web cannot be loaded directly by editors such as
> Proteg� or NeOnToolkit. There is a problem with importing the
should be able to dereference the given PURL to an rdf/xml media typed
document (I tried it with curl and accepted = rdf/xml and it worked).
AFAIK, this is an know issue of Proteg�. You may give TopBraid Composer
a try ;)
> I've found a workaround in the editors I use,
> this is, to load firstly by hand the similarity ontology in the
> workspace and then load the Music Ontology in the same workspace (this
> way, when it imports the namespace, it takes it from the loaded
> version).
>
> Besides, a question has come into my mind once I've managed to load
> it: is this ontology oriented to perform any reasoning task or just to
> specify a vocabulary for publishing Linked Data? I mean, once loaded,
> I've classified it with three different reasoners (Pellet, HermiT and
> Fact++) and all of them tell me that Association (the class from
> similarity ontology) is equivalent to Thing, and therefore, everything
> in the ontology has to be considered an association.
Hm, that's a bit strange. I'm not really sure, how the reasoners came to
this conclusion. From having a look at the MuSim specification, I
couldn't find this equivalence-class relation. There is only a
equivalence relation to owl:Restriction (which is right (AFAIK)).
Cheers,
Bo
thanks a lot for you interest in the Music Ontology.
Here is only a short answer from me for now:
(see my inline comments)
On 07/27/2012 03:52 PM, Carlos Bobed wrote:
> Hello everyone!
>
> I was studying the Music Ontology to use the vocabulary it provides
> and I've found out that the current version:
>
> http://motools.sourceforge.net/doc/musicontology.rdfs
>
> as it is in the Web cannot be loaded directly by editors such as
> similarity ontology (MuSim): there is an ontology that imports it and
> references its namespace http://purl.org/ontology/similarity/ instead
> of the physical URL.
That's probably an issue of your ontology modelling tool, i.e., it
> references its namespace http://purl.org/ontology/similarity/ instead
> of the physical URL.
should be able to dereference the given PURL to an rdf/xml media typed
document (I tried it with curl and accepted = rdf/xml and it worked).
AFAIK, this is an know issue of Proteg�. You may give TopBraid Composer
a try ;)
> I've found a workaround in the editors I use,
> this is, to load firstly by hand the similarity ontology in the
> workspace and then load the Music Ontology in the same workspace (this
> way, when it imports the namespace, it takes it from the loaded
> version).
>
> Besides, a question has come into my mind once I've managed to load
> it: is this ontology oriented to perform any reasoning task or just to
> specify a vocabulary for publishing Linked Data? I mean, once loaded,
> I've classified it with three different reasoners (Pellet, HermiT and
> Fact++) and all of them tell me that Association (the class from
> similarity ontology) is equivalent to Thing, and therefore, everything
> in the ontology has to be considered an association.
this conclusion. From having a look at the MuSim specification, I
couldn't find this equivalence-class relation. There is only a
equivalence relation to owl:Restriction (which is right (AFAIK)).
Cheers,
Bo
Carlos Bobed
Jul 27, 2012, 10:40:05 AM7/27/12
to music-ontology-sp...@googlegroups.com
2012/7/27 Bob Ferris <za...@smiy.org>:
I'll try TopBraid :), thank you. However, this also happens when
loading it via the OWLAPI (at least, version 3.2.x), which
is the API I use to handle ontologies.
>
>> I've found a workaround in the editors I use,
>> this is, to load firstly by hand the similarity ontology in the
>> workspace and then load the Music Ontology in the same workspace (this
>> way, when it imports the namespace, it takes it from the loaded
>> version).
>>
>> Besides, a question has come into my mind once I've managed to load
>> it: is this ontology oriented to perform any reasoning task or just to
>> specify a vocabulary for publishing Linked Data? I mean, once loaded,
>> I've classified it with three different reasoners (Pellet, HermiT and
>> Fact++) and all of them tell me that Association (the class from
>> similarity ontology) is equivalent to Thing, and therefore, everything
>> in the ontology has to be considered an association.
>
>
> Hm, that's a bit strange. I'm not really sure, how the reasoners came to
> this conclusion. From having a look at the MuSim specification, I couldn't
> find this equivalence-class relation. There is only a equivalence relation
> to owl:Restriction (which is right (AFAIK)).
>
> Cheers,
>
it's quite counter-intuitive, but when loading and classifying the
similarity ontology, all the
different reasoners entail that they are equivalent. In particular,
when asking for explanations, Pellet
tells that :
1) Association is equivalent to method max 1 Thing
2) method Domain Association
I think this is related to OWA, and the fact that the properties in
OWL are not "inherited" in the same sense
that in Object Oriented thinking. When we say that the domain of
method is Association, is works in both ways,
if you assert that something participates in the method property, it
becames an Association. This way, the first
fact says that Associations are the elements that participates at most
in method property once: every element that
does not participate at all in method property, will be classified as
Annotation. Moreover, due to the fact of OWA, if
an element is associated to two different individuals through method,
they become the same individual (as nobody has told
the reasoner to assume them as being different individuals).
For example, if I assert a, b, c as being Thing:
- if I assert a method b, then
a and c are Association, b becomes AssociationMethod (due to the range
of method)
- if I add a method c, then
a is an Association, and b and c become equivalent individuals
- now I add an allDifferent axiom to close the world:
the ontology becomes inconsistent.
So, I think there is a problem with the underlying logic that has been
used to model the similarity ontology, that is propagated
to the whole Music Ontology.
Cheers,
> Hi Carlos,
>
> thanks a lot for you interest in the Music Ontology.
>
> Here is only a short answer from me for now:
>
> (see my inline comments)
>
> On 07/27/2012 03:52 PM, Carlos Bobed wrote:
>>
>> Hello everyone!
>>
>> I was studying the Music Ontology to use the vocabulary it provides
>> and I've found out that the current version:
>>
>> http://motools.sourceforge.net/doc/musicontology.rdfs
>>
>> as it is in the Web cannot be loaded directly by editors such as
>> Protegé or NeOnToolkit. There is a problem with importing the
>
> thanks a lot for you interest in the Music Ontology.
>
> Here is only a short answer from me for now:
>
> (see my inline comments)
>
> On 07/27/2012 03:52 PM, Carlos Bobed wrote:
>>
>> Hello everyone!
>>
>> I was studying the Music Ontology to use the vocabulary it provides
>> and I've found out that the current version:
>>
>> http://motools.sourceforge.net/doc/musicontology.rdfs
>>
>> as it is in the Web cannot be loaded directly by editors such as
>> similarity ontology (MuSim): there is an ontology that imports it and
>> references its namespace http://purl.org/ontology/similarity/ instead
>> of the physical URL.
>
>
> That's probably an issue of your ontology modelling tool, i.e., it should be
> able to dereference the given PURL to an rdf/xml media typed document (I
> tried it with curl and accepted = rdf/xml and it worked). AFAIK, this is an
> know issue of Protegé. You may give TopBraid Composer a try ;)
>> references its namespace http://purl.org/ontology/similarity/ instead
>> of the physical URL.
>
>
> That's probably an issue of your ontology modelling tool, i.e., it should be
> able to dereference the given PURL to an rdf/xml media typed document (I
> tried it with curl and accepted = rdf/xml and it worked). AFAIK, this is an
I'll try TopBraid :), thank you. However, this also happens when
loading it via the OWLAPI (at least, version 3.2.x), which
is the API I use to handle ontologies.
>
>> I've found a workaround in the editors I use,
>> this is, to load firstly by hand the similarity ontology in the
>> workspace and then load the Music Ontology in the same workspace (this
>> way, when it imports the namespace, it takes it from the loaded
>> version).
>>
>> Besides, a question has come into my mind once I've managed to load
>> it: is this ontology oriented to perform any reasoning task or just to
>> specify a vocabulary for publishing Linked Data? I mean, once loaded,
>> I've classified it with three different reasoners (Pellet, HermiT and
>> Fact++) and all of them tell me that Association (the class from
>> similarity ontology) is equivalent to Thing, and therefore, everything
>> in the ontology has to be considered an association.
>
>
> Hm, that's a bit strange. I'm not really sure, how the reasoners came to
> this conclusion. From having a look at the MuSim specification, I couldn't
> find this equivalence-class relation. There is only a equivalence relation
> to owl:Restriction (which is right (AFAIK)).
>
> Cheers,
>
similarity ontology, all the
different reasoners entail that they are equivalent. In particular,
when asking for explanations, Pellet
tells that :
1) Association is equivalent to method max 1 Thing
2) method Domain Association
I think this is related to OWA, and the fact that the properties in
OWL are not "inherited" in the same sense
that in Object Oriented thinking. When we say that the domain of
method is Association, is works in both ways,
if you assert that something participates in the method property, it
becames an Association. This way, the first
fact says that Associations are the elements that participates at most
in method property once: every element that
does not participate at all in method property, will be classified as
Annotation. Moreover, due to the fact of OWA, if
an element is associated to two different individuals through method,
they become the same individual (as nobody has told
the reasoner to assume them as being different individuals).
For example, if I assert a, b, c as being Thing:
- if I assert a method b, then
a and c are Association, b becomes AssociationMethod (due to the range
of method)
- if I add a method c, then
a is an Association, and b and c become equivalent individuals
- now I add an allDifferent axiom to close the world:
the ontology becomes inconsistent.
So, I think there is a problem with the underlying logic that has been
used to model the similarity ontology, that is propagated
to the whole Music Ontology.
Cheers,
Carlos Bobed
Jul 27, 2012, 10:48:32 AM7/27/12
to music-ontology-sp...@googlegroups.com
In fact, what the ontology is saying is that:
Everything is an association, because:
1) they are not participating in the method property
2) if they participate, they only can be related to one element.
Which involves every individual, so it is equivalent to Thing.
I think that the Association definition should include also an
existential/minimal cardinality of 1 for the method property in order
to be
accurate.
2012/7/27 Carlos Bobed <carlos...@gmail.com>:
Everything is an association, because:
1) they are not participating in the method property
2) if they participate, they only can be related to one element.
Which involves every individual, so it is equivalent to Thing.
I think that the Association definition should include also an
existential/minimal cardinality of 1 for the method property in order
to be
accurate.
2012/7/27 Carlos Bobed <carlos...@gmail.com>:
Sean Bechhofer
Jul 27, 2012, 11:56:42 AM7/27/12
to music-ontology-sp...@googlegroups.com
On 27 Jul 2012, at 15:48, Carlos Bobed wrote:
> In fact, what the ontology is saying is that:
>
> Everything is an association, because:
>
> 1) they are not participating in the method property
> 2) if they participate, they only can be related to one element.
Association EquivalentTo method max 1 Thing
method Domain Association
So anything that doesn't have a method relationship has less than one method relationship (i.e 0), so is thus an Association. And anything with a method relationship must be an Association because of the domain assertion. Therefore everything is an Association, regardless of whether we know about its method relationships, which is where open world reasoning comes in.
Cheers,
Sean
> You received this message because you are subscribed to the Google Groups "Music Ontology Specification Group" group.
> To post to this group, send email to music-ontology-sp...@googlegroups.com.
> To unsubscribe from this group, send email to music-ontology-specific...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/music-ontology-specification-group?hl=en.
>
--
sean.be...@googlemail.com
Yves Raimond
Jul 27, 2012, 12:10:52 PM7/27/12
to music-ontology-sp...@googlegroups.com
Hello!
I just forwarded this discussion to Kurt, as I think musim is under his control - it looks like we might need to remove this axiom from there?
In the short term, perhaps we can remove the import statement?
Best
Yves
Frederick Giasson
Jul 27, 2012, 12:30:23 PM7/27/12
to music-ontology-sp...@googlegroups.com
Hi!
> In the short term, perhaps we can remove the import statement?
>
Agreed
Thanks,
Fred
> In the short term, perhaps we can remove the import statement?
>
Thanks,
Fred
Bob Ferris
Jul 28, 2012, 12:47:16 PM7/28/12
to music-ontology-sp...@googlegroups.com
Hi Carlos,
Hi all,
now I see the problem, the modelling of the restriction is wrong. It
should be a sub class of owl:Restriction, instead of a equivalent class.
So thanks a lot for discovering the bug ;)
Cheers,
Bo
On 7/27/2012 4:40 PM, Carlos Bobed wrote:
> 2012/7/27 Bob Ferris <za...@smiy.org>:
>> Hi Carlos,
>>
>> thanks a lot for you interest in the Music Ontology.
>>
>> Here is only a short answer from me for now:
>>
>> (see my inline comments)
>>
>> On 07/27/2012 03:52 PM, Carlos Bobed wrote:
>>>
>>> Hello everyone!
>>>
>>> I was studying the Music Ontology to use the vocabulary it provides
>>> and I've found out that the current version:
>>>
>>> http://motools.sourceforge.net/doc/musicontology.rdfs
>>>
>>> as it is in the Web cannot be loaded directly by editors such as
>>> Proteg� or NeOnToolkit. There is a problem with importing the
Hi all,
now I see the problem, the modelling of the restriction is wrong. It
should be a sub class of owl:Restriction, instead of a equivalent class.
So thanks a lot for discovering the bug ;)
Cheers,
Bo
On 7/27/2012 4:40 PM, Carlos Bobed wrote:
> 2012/7/27 Bob Ferris <za...@smiy.org>:
>> Hi Carlos,
>>
>> thanks a lot for you interest in the Music Ontology.
>>
>> Here is only a short answer from me for now:
>>
>> (see my inline comments)
>>
>> On 07/27/2012 03:52 PM, Carlos Bobed wrote:
>>>
>>> Hello everyone!
>>>
>>> I was studying the Music Ontology to use the vocabulary it provides
>>> and I've found out that the current version:
>>>
>>> http://motools.sourceforge.net/doc/musicontology.rdfs
>>>
>>> as it is in the Web cannot be loaded directly by editors such as
>>> similarity ontology (MuSim): there is an ontology that imports it and
>>> references its namespace http://purl.org/ontology/similarity/ instead
>>> of the physical URL.
>>
>>
>> That's probably an issue of your ontology modelling tool, i.e., it should be
>> able to dereference the given PURL to an rdf/xml media typed document (I
>> tried it with curl and accepted = rdf/xml and it worked). AFAIK, this is an
>> know issue of Proteg�. You may give TopBraid Composer a try ;)
>>> references its namespace http://purl.org/ontology/similarity/ instead
>>> of the physical URL.
>>
>>
>> That's probably an issue of your ontology modelling tool, i.e., it should be
>> able to dereference the given PURL to an rdf/xml media typed document (I
>> tried it with curl and accepted = rdf/xml and it worked). AFAIK, this is an
Carlos Bobed
Jul 28, 2012, 3:20:20 PM7/28/12
to music-ontology-sp...@googlegroups.com
Hi Bob,
Hi all,
You're welcome, the community will profit :)
El 28/07/2012 17:47, "Bob Ferris" <za...@smiy.org> escribió:
Hi Carlos,
Hi all,
now I see the problem, the modelling of the restriction is wrong. It should be a sub class of owl:Restriction, instead of a equivalent class. So thanks a lot for discovering the bug ;)
Cheers,
Bo
On 7/27/2012 4:40 PM, Carlos Bobed wrote:
2012/7/27 Bob Ferris <za...@smiy.org>:
Hi Carlos,
thanks a lot for you interest in the Music Ontology.
Here is only a short answer from me for now:
(see my inline comments)
On 07/27/2012 03:52 PM, Carlos Bobed wrote:
Hello everyone!
I was studying the Music Ontology to use the vocabulary it provides
and I've found out that the current version:
http://motools.sourceforge.net/doc/musicontology.rdfs
as it is in the Web cannot be loaded directly by editors such as
Protegé or NeOnToolkit. There is a problem with importing the
similarity ontology (MuSim): there is an ontology that imports it and
references its namespace http://purl.org/ontology/similarity/ instead
of the physical URL.
That's probably an issue of your ontology modelling tool, i.e., it should be
able to dereference the given PURL to an rdf/xml media typed document (I
tried it with curl and accepted = rdf/xml and it worked). AFAIK, this is an
know issue of Protegé. You may give TopBraid Composer a try ;)
--
You received this message because you are subscribed to the Google Groups "Music Ontology Specification Group" group.
To post to this group, send email to music-ontology-specification-gr...@googlegroups.com.
To unsubscribe from this group, send email to music-ontology-specification-group+unsubscribe@googlegroups.com.
Reply all
Reply to author
Forward
0 new messages