FW: advice on adding info to other people's ontology

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

FW: advice on adding info to other people's ontology

Lieke Verhelst

Hi all,

Adding statements to an entity from a namespace that is not yours the way
Michel suggests is saying things about someone else's things like you
were that other party.  It is assuming that this is technically possible.
( It isn't, because you cannot publish the statement in someone else's
namespace since it is on someone else's server.) 
The editors allow it but imho this type of practice must be avoided
because it causes unexpected results, just like Ron says. I haven't
checked but hopefully reasoners act on the real origin of statements on
the web and ignore the statements about remote namespaces that are locally
declared.

Of course it is perfectly fine to add statements in your own namespace
that say something about someone else's entities.


Best, Lieke
Linked Data Factory
Semantic web and Linked Data Services
The Netherlands

-----Original Message-----
From: Bohms, H.M. (Michel) [mailto:[hidden email]] 
Sent: 29 April 2014 19:11
To: [hidden email]'; 'User support for WebProtege and Protege
Desktop'
Cc: 'Lieke Verhelst'
Subject: RE: [protege-user] advice on adding info to other people's
ontology

Great, so I actually use it in the spirit of.
I'll keep it the way I have it (with the triple added on my side)

Thx Joshua for your clear analysis, Michel

-----Original Message-----
From: protege-user [mailto:[hidden email]] On
Behalf Of Joshua TAYLOR
Sent: dinsdag 29 april 2014 17:19
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] advice on adding info to other people's
ontology

On Tue, Apr 29, 2014 at 11:15 AM, Joshua TAYLOR <[hidden email]>
wrote:
If you have a property op_hasPart that's a subproperty of some DC property, then the DC property will still need to be declared as either an annotation property, a datatype property, or an object property, and the only one that makes sense here would be as an object property. So this doesn't really avoid the problem.
it is worth noting though, that the documentation about hasPart [1]
says: "This term is intended to be used with non-literal values as defined
in the DCMI Abstract Model
(http://dublincore.org/documents/abstract-model/). As of December 2007,
the DCMI Usage Board is seeking a way to express this intention with a
formal range declaration."  It doesn't have a formal range, but the fact
that it's intended to be used with non-literal values at least means that
a declaration as an owl:ObjectProperty won't be too far from its intended
use.  If nothing else, the values it will have as an object property will
be non-literals, so any triples with the owl:ObjectProperty hasParts will
be compatible the rdf:Property hasPart.

//JT

[1] http://dublincore.org/documents/dcmi-terms/#terms-hasPart

--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user



Dit bericht kan informatie bevatten die niet voor u is bestemd. Indien u
niet de geadresseerde bent of dit bericht abusievelijk aan u is
toegezonden, wordt u verzocht dat aan de afzender te melden en het bericht
te verwijderen. TNO aanvaardt geen aansprakelijkheid voor de inhoud van
deze e-mail, de wijze waarop u deze gebruikt en voor schade, van welke
aard ook, die verband houdt met risico's verbonden aan het elektronisch
verzenden van berichten.

 

This message may contain information that is not intended for you. If you
are not the addressee or if this message was sent to you by mistake, you
are requested to inform the sender and delete the message. TNO accepts no
liability for the content of this e-mail, for the manner in which you use
it and for damage of any kind resulting from the risks inherent to the
electronic transmission of messages.


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: FW: advice on adding info to other people's ontology

Bohms, H.M. (Michel)

One more try after >

 

 

Dr. ir. H.M. (Michel) Bohms
Sr. Research Scientist
Structural Reliability

T +31 (0)88 866 31 07
M +31 (0)63 038 12 20
[hidden email]

Location

 

This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.

 

From: protege-user [mailto:[hidden email]] On Behalf Of Lieke Verhelst
Sent: woensdag 30 april 2014 15:03
To: [hidden email]
Subject: [protege-user] FW: advice on adding info to other people's ontology

 

 

Hi all,
 
Adding statements to an entity from a namespace that is not yours the way
Michel suggests is saying things about someone else's things like you
were that other party.  
 
>no, not like “being that other party” Just me. I add the triple in my namespace.
 
 
It is assuming that this is technically possible.
 
>no, not assuming, it IS possible.
 
 
( It isn't, because you cannot publish the statement in someone else's
namespace since it is on someone else's server.) 
 
>yes, it is, since I am NOT doing that in someone elses name space but in my own name space..
(I just adapt so this could hurt interoperability; but not in this case because my change is in the spirit of the DC intention…)
 
 
The editors allow it but imho this type of practice must be avoided
because it causes unexpected results, just like Ron says. I haven't
checked but hopefully reasoners act on the real origin of statements on
the web and ignore the statements about remote namespaces that are locally
declared.
 
>really, I do not change stuff in other people’s name space. Even if it were technically possible I would not there. People are the bos over their own name space.
 
Of course it is perfectly fine to add statements in your own namespace
that say something about someone else's entities.
 
>right, and this is what I do/did.
 
 
Best, Lieke
Linked Data Factory
Semantic web and Linked Data Services
The Netherlands
 
-----Original Message-----
From: Bohms, H.M. (Michel) [mailto:[hidden email]] 
Sent: 29 April 2014 19:11
To: [hidden email]'; 'User support for WebProtege and Protege
Desktop'
Cc: 'Lieke Verhelst'
Subject: RE: [protege-user] advice on adding info to other people's
ontology
 
Great, so I actually use it in the spirit of.
I'll keep it the way I have it (with the triple added on my side)
 
Thx Joshua for your clear analysis, Michel
 
-----Original Message-----
From: protege-user [mailto:[hidden email]] On
Behalf Of Joshua TAYLOR
Sent: dinsdag 29 april 2014 17:19
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] advice on adding info to other people's
ontology
 
On Tue, Apr 29, 2014 at 11:15 AM, Joshua TAYLOR <[hidden email]>
wrote:

If you have a property op_hasPart that's a subproperty of some DC property, then the DC property will still need to be declared as either an annotation property, a datatype property, or an object property, and the only one that makes sense here would be as an object property. So this doesn't really avoid the problem.

it is worth noting though, that the documentation about hasPart [1]
says: "This term is intended to be used with non-literal values as defined
in the DCMI Abstract Model
(http://dublincore.org/documents/abstract-model/). As of December 2007,
the DCMI Usage Board is seeking a way to express this intention with a
formal range declaration."  It doesn't have a formal range, but the fact
that it's intended to be used with non-literal values at least means that
a declaration as an owl:ObjectProperty won't be too far from its intended
use.  If nothing else, the values it will have as an object property will
be non-literals, so any triples with the owl:ObjectProperty hasParts will
be compatible the rdf:Property hasPart.
 
//JT
 
[1] http://dublincore.org/documents/dcmi-terms/#terms-hasPart
 
--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
 
 
 
Dit bericht kan informatie bevatten die niet voor u is bestemd. Indien u
niet de geadresseerde bent of dit bericht abusievelijk aan u is
toegezonden, wordt u verzocht dat aan de afzender te melden en het bericht
te verwijderen. TNO aanvaardt geen aansprakelijkheid voor de inhoud van
deze e-mail, de wijze waarop u deze gebruikt en voor schade, van welke
aard ook, die verband houdt met risico's verbonden aan het elektronisch
verzenden van berichten.
 
 
 
This message may contain information that is not intended for you. If you
are not the addressee or if this message was sent to you by mistake, you
are requested to inform the sender and delete the message. TNO accepts no
liability for the content of this e-mail, for the manner in which you use
it and for damage of any kind resulting from the risks inherent to the
electronic transmission of messages.
 

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: FW: advice on adding info to other people's ontology

Ron Rudnicki
In reply to this post by Lieke Verhelst

But consider. I take your class: Person and to it I add an axiom that Person is a subclass of things that wear hats. Now I publish my ontology claiming that the class person is the same as yours. I also collect instance data about Persons using my class. You try to aggregrate my data with yours, funny thing is that all these people wear hats…

 

What I should have done instead was take your class as a parent class and create a subclass called PersonWithHat or HattedPerson. Then there’s no problem.

 

From: protege-user [mailto:[hidden email]] On Behalf Of Lieke Verhelst
Sent: Wednesday, April 30, 2014 9:03 AM
To: [hidden email]
Subject: [protege-user] FW: advice on adding info to other people's ontology

 

 

Hi all,
 
Adding statements to an entity from a namespace that is not yours the way
Michel suggests is saying things about someone else's things like you
were that other party.  It is assuming that this is technically possible.
( It isn't, because you cannot publish the statement in someone else's
namespace since it is on someone else's server.) 
The editors allow it but imho this type of practice must be avoided
because it causes unexpected results, just like Ron says. I haven't
checked but hopefully reasoners act on the real origin of statements on
the web and ignore the statements about remote namespaces that are locally
declared.
 
Of course it is perfectly fine to add statements in your own namespace
that say something about someone else's entities.
 
 
Best, Lieke
Linked Data Factory
Semantic web and Linked Data Services
The Netherlands
 
-----Original Message-----
From: Bohms, H.M. (Michel) [mailto:[hidden email]] 
Sent: 29 April 2014 19:11
To: [hidden email]'; 'User support for WebProtege and Protege
Desktop'
Cc: 'Lieke Verhelst'
Subject: RE: [protege-user] advice on adding info to other people's
ontology
 
Great, so I actually use it in the spirit of.
I'll keep it the way I have it (with the triple added on my side)
 
Thx Joshua for your clear analysis, Michel
 
-----Original Message-----
From: protege-user [mailto:[hidden email]] On
Behalf Of Joshua TAYLOR
Sent: dinsdag 29 april 2014 17:19
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] advice on adding info to other people's
ontology
 
On Tue, Apr 29, 2014 at 11:15 AM, Joshua TAYLOR <[hidden email]>
wrote:

If you have a property op_hasPart that's a subproperty of some DC property, then the DC property will still need to be declared as either an annotation property, a datatype property, or an object property, and the only one that makes sense here would be as an object property. So this doesn't really avoid the problem.

it is worth noting though, that the documentation about hasPart [1]
says: "This term is intended to be used with non-literal values as defined
in the DCMI Abstract Model
(http://dublincore.org/documents/abstract-model/). As of December 2007,
the DCMI Usage Board is seeking a way to express this intention with a
formal range declaration."  It doesn't have a formal range, but the fact
that it's intended to be used with non-literal values at least means that
a declaration as an owl:ObjectProperty won't be too far from its intended
use.  If nothing else, the values it will have as an object property will
be non-literals, so any triples with the owl:ObjectProperty hasParts will
be compatible the rdf:Property hasPart.
 
//JT
 
[1] http://dublincore.org/documents/dcmi-terms/#terms-hasPart
 
--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
 
 
 
Dit bericht kan informatie bevatten die niet voor u is bestemd. Indien u
niet de geadresseerde bent of dit bericht abusievelijk aan u is
toegezonden, wordt u verzocht dat aan de afzender te melden en het bericht
te verwijderen. TNO aanvaardt geen aansprakelijkheid voor de inhoud van
deze e-mail, de wijze waarop u deze gebruikt en voor schade, van welke
aard ook, die verband houdt met risico's verbonden aan het elektronisch
verzenden van berichten.
 
 
 
This message may contain information that is not intended for you. If you
are not the addressee or if this message was sent to you by mistake, you
are requested to inform the sender and delete the message. TNO accepts no
liability for the content of this e-mail, for the manner in which you use
it and for damage of any kind resulting from the risks inherent to the
electronic transmission of messages.
 

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: FW: advice on adding info to other people's ontology

Bohms, H.M. (Michel)

I assumed it works in MY/THIS case because I added a triple that was actually in consideration of / in the spirit of the semantics intended by DCI…(see argument Joshua)

(so not being diff. semantics like in your example with and without hats)

 

So what you say is:

Make a new own hasPart being a subclass of dcterms:haPart and then add the fact that my hasPart is of type owl:objectProperty

To be even more cleaner and independent of my assumption that I actually model what was probably meant by DCI…

 

???

 

 

Dr. ir. H.M. (Michel) Bohms
Sr. Research Scientist
Structural Reliability

T +31 (0)88 866 31 07
M +31 (0)63 038 12 20
[hidden email]

Location

 

This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.

 

From: protege-user [mailto:[hidden email]] On Behalf Of Ron Rudnicki
Sent: woensdag 30 april 2014 15:23
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] FW: advice on adding info to other people's ontology

 

But consider. I take your class: Person and to it I add an axiom that Person is a subclass of things that wear hats. Now I publish my ontology claiming that the class person is the same as yours. I also collect instance data about Persons using my class. You try to aggregrate my data with yours, funny thing is that all these people wear hats…

 

What I should have done instead was take your class as a parent class and create a subclass called PersonWithHat or HattedPerson. Then there’s no problem.

 

From: protege-user [[hidden email]] On Behalf Of Lieke Verhelst
Sent: Wednesday, April 30, 2014 9:03 AM
To: [hidden email]
Subject: [protege-user] FW: advice on adding info to other people's ontology

 

 

Hi all,
 
Adding statements to an entity from a namespace that is not yours the way
Michel suggests is saying things about someone else's things like you
were that other party.  It is assuming that this is technically possible.
( It isn't, because you cannot publish the statement in someone else's
namespace since it is on someone else's server.) 
The editors allow it but imho this type of practice must be avoided
because it causes unexpected results, just like Ron says. I haven't
checked but hopefully reasoners act on the real origin of statements on
the web and ignore the statements about remote namespaces that are locally
declared.
 
Of course it is perfectly fine to add statements in your own namespace
that say something about someone else's entities.
 
 
Best, Lieke
Linked Data Factory
Semantic web and Linked Data Services
The Netherlands
 
-----Original Message-----
From: Bohms, H.M. (Michel) [mailto:[hidden email]] 
Sent: 29 April 2014 19:11
To: [hidden email]'; 'User support for WebProtege and Protege
Desktop'
Cc: 'Lieke Verhelst'
Subject: RE: [protege-user] advice on adding info to other people's
ontology
 
Great, so I actually use it in the spirit of.
I'll keep it the way I have it (with the triple added on my side)
 
Thx Joshua for your clear analysis, Michel
 
-----Original Message-----
From: protege-user [mailto:[hidden email]] On
Behalf Of Joshua TAYLOR
Sent: dinsdag 29 april 2014 17:19
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] advice on adding info to other people's
ontology
 
On Tue, Apr 29, 2014 at 11:15 AM, Joshua TAYLOR <[hidden email]>
wrote:

If you have a property op_hasPart that's a subproperty of some DC property, then the DC property will still need to be declared as either an annotation property, a datatype property, or an object property, and the only one that makes sense here would be as an object property. So this doesn't really avoid the problem.

it is worth noting though, that the documentation about hasPart [1]
says: "This term is intended to be used with non-literal values as defined
in the DCMI Abstract Model
(http://dublincore.org/documents/abstract-model/). As of December 2007,
the DCMI Usage Board is seeking a way to express this intention with a
formal range declaration."  It doesn't have a formal range, but the fact
that it's intended to be used with non-literal values at least means that
a declaration as an owl:ObjectProperty won't be too far from its intended
use.  If nothing else, the values it will have as an object property will
be non-literals, so any triples with the owl:ObjectProperty hasParts will
be compatible the rdf:Property hasPart.
 
//JT
 
[1] http://dublincore.org/documents/dcmi-terms/#terms-hasPart
 
--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
 
 
 
Dit bericht kan informatie bevatten die niet voor u is bestemd. Indien u
niet de geadresseerde bent of dit bericht abusievelijk aan u is
toegezonden, wordt u verzocht dat aan de afzender te melden en het bericht
te verwijderen. TNO aanvaardt geen aansprakelijkheid voor de inhoud van
deze e-mail, de wijze waarop u deze gebruikt en voor schade, van welke
aard ook, die verband houdt met risico's verbonden aan het elektronisch
verzenden van berichten.
 
 
 
This message may contain information that is not intended for you. If you
are not the addressee or if this message was sent to you by mistake, you
are requested to inform the sender and delete the message. TNO accepts no
liability for the content of this e-mail, for the manner in which you use
it and for damage of any kind resulting from the risks inherent to the
electronic transmission of messages.
 

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: FW: advice on adding info to other people's ontology

Joshua TAYLOR
In reply to this post by Ron Rudnicki
On Wed, Apr 30, 2014 at 9:23 AM, Ron Rudnicki <[hidden email]> wrote:
> What I should have done instead was take your class as a parent class and
> create a subclass called PersonWithHat or HattedPerson. Then there’s no
> problem.

With OWL ontologies, this is a better approach for interoperability.
Note that it doesn't have anything to do with *where* data is
published;  it's just a matter of what axioms someone it using.  If
you use axioms about an entity that don't come from "the authoritative
source" about that entity, it shouldn't be surprising if
incompatibilities arise.  The ability to say things about something
"in another namespace" is one of the big *benefits* of using URIs as
identifiers.  (I use scare quotes around "in another namespace"
because the whole concept of namespace is really just convention.  The
fact of the matter is that we use URIs for identifiers. Namespaces
only arise because it's easier to deal with a bunch of URIs that have
something textually in common.  There's nothing about the textual
content of the URI that affects its meaning logically.

The main point, though, was that "With OWL ontologies, this is a
better approach for interoperability."  The problem in this case is
that the DC terms vocabulary *isn't* an OWL ontology.  This means that
dcterms:subject (or dc:subject, or whatever started this conversation)
isn't declared by its creators as any type of OWL property, so there's
no ontology to import that would provide that declaration.  Thus, even
to define some property as a subproperty of it would require first
declaring it as an OWL property first.

The problem here isn't so much about "what can/ought we do/not do with
'other people's URIs'?" but more about "how do we reuse useful RDF
vocabularies in OWL ontologies?

//JT


--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: FW: advice on adding info to other people's ontology

Ron Rudnicki
In reply to this post by Bohms, H.M. (Michel)

If my caution applies at all, it applies only if DC publishes an .owl file in which dcterms:hasPart is typed as something other than owl:ObjectProperty. If it does, then by changing  the type in your file there could be queries that unexpectedly return results from your ontology but not from another that uses the canonical DC file or vice versa. In this case if you subtype, then you are warning your users that some extra effort will be needed to merge data, thereby saving us some debugging efforts. Of course, If DC offers no guidance on the implementation of hasPart, then typing it for one’s own namespace is both legitimate and necessary.

 

From: protege-user [mailto:[hidden email]] On Behalf Of Bohms, H.M. (Michel)
Sent: Wednesday, April 30, 2014 9:32 AM
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] FW: advice on adding info to other people's ontology

 

I assumed it works in MY/THIS case because I added a triple that was actually in consideration of / in the spirit of the semantics intended by DCI…(see argument Joshua)

(so not being diff. semantics like in your example with and without hats)

 

So what you say is:

Make a new own hasPart being a subclass of dcterms:haPart and then add the fact that my hasPart is of type owl:objectProperty

To be even more cleaner and independent of my assumption that I actually model what was probably meant by DCI…

 

???

 

 

Dr. ir. H.M. (Michel) Bohms
Sr. Research Scientist
Structural Reliability

T +31 (0)88 866 31 07
M +31 (0)63 038 12 20
[hidden email]

Location

 

This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.

 

From: protege-user [[hidden email]] On Behalf Of Ron Rudnicki
Sent: woensdag 30 april 2014 15:23
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] FW: advice on adding info to other people's ontology

 

But consider. I take your class: Person and to it I add an axiom that Person is a subclass of things that wear hats. Now I publish my ontology claiming that the class person is the same as yours. I also collect instance data about Persons using my class. You try to aggregrate my data with yours, funny thing is that all these people wear hats…

 

What I should have done instead was take your class as a parent class and create a subclass called PersonWithHat or HattedPerson. Then there’s no problem.

 

From: protege-user [[hidden email]] On Behalf Of Lieke Verhelst
Sent: Wednesday, April 30, 2014 9:03 AM
To: [hidden email]
Subject: [protege-user] FW: advice on adding info to other people's ontology

 

 

Hi all,
 
Adding statements to an entity from a namespace that is not yours the way
Michel suggests is saying things about someone else's things like you
were that other party.  It is assuming that this is technically possible.
( It isn't, because you cannot publish the statement in someone else's
namespace since it is on someone else's server.) 
The editors allow it but imho this type of practice must be avoided
because it causes unexpected results, just like Ron says. I haven't
checked but hopefully reasoners act on the real origin of statements on
the web and ignore the statements about remote namespaces that are locally
declared.
 
Of course it is perfectly fine to add statements in your own namespace
that say something about someone else's entities.
 
 
Best, Lieke
Linked Data Factory
Semantic web and Linked Data Services
The Netherlands
 
-----Original Message-----
From: Bohms, H.M. (Michel) [mailto:[hidden email]] 
Sent: 29 April 2014 19:11
To: [hidden email]'; 'User support for WebProtege and Protege
Desktop'
Cc: 'Lieke Verhelst'
Subject: RE: [protege-user] advice on adding info to other people's
ontology
 
Great, so I actually use it in the spirit of.
I'll keep it the way I have it (with the triple added on my side)
 
Thx Joshua for your clear analysis, Michel
 
-----Original Message-----
From: protege-user [mailto:[hidden email]] On
Behalf Of Joshua TAYLOR
Sent: dinsdag 29 april 2014 17:19
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] advice on adding info to other people's
ontology
 
On Tue, Apr 29, 2014 at 11:15 AM, Joshua TAYLOR <[hidden email]>
wrote:

If you have a property op_hasPart that's a subproperty of some DC property, then the DC property will still need to be declared as either an annotation property, a datatype property, or an object property, and the only one that makes sense here would be as an object property. So this doesn't really avoid the problem.

it is worth noting though, that the documentation about hasPart [1]
says: "This term is intended to be used with non-literal values as defined
in the DCMI Abstract Model
(http://dublincore.org/documents/abstract-model/). As of December 2007,
the DCMI Usage Board is seeking a way to express this intention with a
formal range declaration."  It doesn't have a formal range, but the fact
that it's intended to be used with non-literal values at least means that
a declaration as an owl:ObjectProperty won't be too far from its intended
use.  If nothing else, the values it will have as an object property will
be non-literals, so any triples with the owl:ObjectProperty hasParts will
be compatible the rdf:Property hasPart.
 
//JT
 
[1] http://dublincore.org/documents/dcmi-terms/#terms-hasPart
 
--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
 
 
 
Dit bericht kan informatie bevatten die niet voor u is bestemd. Indien u
niet de geadresseerde bent of dit bericht abusievelijk aan u is
toegezonden, wordt u verzocht dat aan de afzender te melden en het bericht
te verwijderen. TNO aanvaardt geen aansprakelijkheid voor de inhoud van
deze e-mail, de wijze waarop u deze gebruikt en voor schade, van welke
aard ook, die verband houdt met risico's verbonden aan het elektronisch
verzenden van berichten.
 
 
 
This message may contain information that is not intended for you. If you
are not the addressee or if this message was sent to you by mistake, you
are requested to inform the sender and delete the message. TNO accepts no
liability for the content of this e-mail, for the manner in which you use
it and for damage of any kind resulting from the risks inherent to the
electronic transmission of messages.
 

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: FW: advice on adding info to other people's ontology

Bohms, H.M. (Michel)

Can we summarise (?):

1.      Best way: make 100% sure no issues: subclass own from source or reinvent (without subclassing)

2.      Reasonable way: assume something that is somehow missing in the authorative source which is not contradicted by the source

a.      What would be IN THIS CASE the “unexpected result” by a reasoner?

3.      Do not: assume things that likely contradict with explicit/implicit statements by the source (hats example)

 

This whole discussion does not only aply to dcterms:hasPart but also to the qudt units we reuse.

I assume they also fall in category 2. Since the sources describes several use cases where for each use case you might have to add some stuff yourself first to make it work

(here: add the triple: <http://qudt.org/schema/qudt#Unit> rdfs:subClassOf rdfs:Datatype .)

 

 

 

 

 

 

Dr. ir. H.M. (Michel) Bohms
Sr. Research Scientist
Structural Reliability

T +31 (0)88 866 31 07
M +31 (0)63 038 12 20
[hidden email]

Location

 

This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.

 

From: protege-user [mailto:[hidden email]] On Behalf Of Ron Rudnicki
Sent: woensdag 30 april 2014 16:05
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] FW: advice on adding info to other people's ontology

 

If my caution applies at all, it applies only if DC publishes an .owl file in which dcterms:hasPart is typed as something other than owl:ObjectProperty. If it does, then by changing  the type in your file there could be queries that unexpectedly return results from your ontology but not from another that uses the canonical DC file or vice versa. In this case if you subtype, then you are warning your users that some extra effort will be needed to merge data, thereby saving us some debugging efforts. Of course, If DC offers no guidance on the implementation of hasPart, then typing it for one’s own namespace is both legitimate and necessary.

 

From: protege-user [[hidden email]] On Behalf Of Bohms, H.M. (Michel)
Sent: Wednesday, April 30, 2014 9:32 AM
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] FW: advice on adding info to other people's ontology

 

I assumed it works in MY/THIS case because I added a triple that was actually in consideration of / in the spirit of the semantics intended by DCI…(see argument Joshua)

(so not being diff. semantics like in your example with and without hats)

 

So what you say is:

Make a new own hasPart being a subclass of dcterms:haPart and then add the fact that my hasPart is of type owl:objectProperty

To be even more cleaner and independent of my assumption that I actually model what was probably meant by DCI…

 

???

 

 

Dr. ir. H.M. (Michel) Bohms
Sr. Research Scientist
Structural Reliability

T +31 (0)88 866 31 07
M +31 (0)63 038 12 20
[hidden email]

Location

 

This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.

 

From: protege-user [[hidden email]] On Behalf Of Ron Rudnicki
Sent: woensdag 30 april 2014 15:23
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] FW: advice on adding info to other people's ontology

 

But consider. I take your class: Person and to it I add an axiom that Person is a subclass of things that wear hats. Now I publish my ontology claiming that the class person is the same as yours. I also collect instance data about Persons using my class. You try to aggregrate my data with yours, funny thing is that all these people wear hats…

 

What I should have done instead was take your class as a parent class and create a subclass called PersonWithHat or HattedPerson. Then there’s no problem.

 

From: protege-user [[hidden email]] On Behalf Of Lieke Verhelst
Sent: Wednesday, April 30, 2014 9:03 AM
To: [hidden email]
Subject: [protege-user] FW: advice on adding info to other people's ontology

 

 

Hi all,
 
Adding statements to an entity from a namespace that is not yours the way
Michel suggests is saying things about someone else's things like you
were that other party.  It is assuming that this is technically possible.
( It isn't, because you cannot publish the statement in someone else's
namespace since it is on someone else's server.) 
The editors allow it but imho this type of practice must be avoided
because it causes unexpected results, just like Ron says. I haven't
checked but hopefully reasoners act on the real origin of statements on
the web and ignore the statements about remote namespaces that are locally
declared.
 
Of course it is perfectly fine to add statements in your own namespace
that say something about someone else's entities.
 
 
Best, Lieke
Linked Data Factory
Semantic web and Linked Data Services
The Netherlands
 
-----Original Message-----
From: Bohms, H.M. (Michel) [mailto:[hidden email]] 
Sent: 29 April 2014 19:11
To: [hidden email]'; 'User support for WebProtege and Protege
Desktop'
Cc: 'Lieke Verhelst'
Subject: RE: [protege-user] advice on adding info to other people's
ontology
 
Great, so I actually use it in the spirit of.
I'll keep it the way I have it (with the triple added on my side)
 
Thx Joshua for your clear analysis, Michel
 
-----Original Message-----
From: protege-user [mailto:[hidden email]] On
Behalf Of Joshua TAYLOR
Sent: dinsdag 29 april 2014 17:19
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] advice on adding info to other people's
ontology
 
On Tue, Apr 29, 2014 at 11:15 AM, Joshua TAYLOR <[hidden email]>
wrote:

If you have a property op_hasPart that's a subproperty of some DC property, then the DC property will still need to be declared as either an annotation property, a datatype property, or an object property, and the only one that makes sense here would be as an object property. So this doesn't really avoid the problem.

it is worth noting though, that the documentation about hasPart [1]
says: "This term is intended to be used with non-literal values as defined
in the DCMI Abstract Model
(http://dublincore.org/documents/abstract-model/). As of December 2007,
the DCMI Usage Board is seeking a way to express this intention with a
formal range declaration."  It doesn't have a formal range, but the fact
that it's intended to be used with non-literal values at least means that
a declaration as an owl:ObjectProperty won't be too far from its intended
use.  If nothing else, the values it will have as an object property will
be non-literals, so any triples with the owl:ObjectProperty hasParts will
be compatible the rdf:Property hasPart.
 
//JT
 
[1] http://dublincore.org/documents/dcmi-terms/#terms-hasPart
 
--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
 
 
 
Dit bericht kan informatie bevatten die niet voor u is bestemd. Indien u
niet de geadresseerde bent of dit bericht abusievelijk aan u is
toegezonden, wordt u verzocht dat aan de afzender te melden en het bericht
te verwijderen. TNO aanvaardt geen aansprakelijkheid voor de inhoud van
deze e-mail, de wijze waarop u deze gebruikt en voor schade, van welke
aard ook, die verband houdt met risico's verbonden aan het elektronisch
verzenden van berichten.
 
 
 
This message may contain information that is not intended for you. If you
are not the addressee or if this message was sent to you by mistake, you
are requested to inform the sender and delete the message. TNO accepts no
liability for the content of this e-mail, for the manner in which you use
it and for damage of any kind resulting from the risks inherent to the
electronic transmission of messages.
 

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: FW: advice on adding info to other people's ontology

Ron Rudnicki

I think this is a good summary.

 

From: protege-user [mailto:[hidden email]] On Behalf Of Bohms, H.M. (Michel)
Sent: Wednesday, April 30, 2014 10:30 AM
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] FW: advice on adding info to other people's ontology

 

Can we summarise (?):

1.       Best way: make 100% sure no issues: subclass own from source or reinvent (without subclassing)

2.       Reasonable way: assume something that is somehow missing in the authorative source which is not contradicted by the source

a.       What would be IN THIS CASE the “unexpected result” by a reasoner?

3.       Do not: assume things that likely contradict with explicit/implicit statements by the source (hats example)

 

This whole discussion does not only aply to dcterms:hasPart but also to the qudt units we reuse.

I assume they also fall in category 2. Since the sources describes several use cases where for each use case you might have to add some stuff yourself first to make it work

(here: add the triple: <http://qudt.org/schema/qudt#Unit> rdfs:subClassOf rdfs:Datatype .)

 

 

 

 

 

 

Dr. ir. H.M. (Michel) Bohms
Sr. Research Scientist
Structural Reliability

T +31 (0)88 866 31 07
M +31 (0)63 038 12 20
[hidden email]

Location

 

This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.

 

From: protege-user [[hidden email]] On Behalf Of Ron Rudnicki
Sent: woensdag 30 april 2014 16:05
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] FW: advice on adding info to other people's ontology

 

If my caution applies at all, it applies only if DC publishes an .owl file in which dcterms:hasPart is typed as something other than owl:ObjectProperty. If it does, then by changing  the type in your file there could be queries that unexpectedly return results from your ontology but not from another that uses the canonical DC file or vice versa. In this case if you subtype, then you are warning your users that some extra effort will be needed to merge data, thereby saving us some debugging efforts. Of course, If DC offers no guidance on the implementation of hasPart, then typing it for one’s own namespace is both legitimate and necessary.

 

From: protege-user [[hidden email]] On Behalf Of Bohms, H.M. (Michel)
Sent: Wednesday, April 30, 2014 9:32 AM
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] FW: advice on adding info to other people's ontology

 

I assumed it works in MY/THIS case because I added a triple that was actually in consideration of / in the spirit of the semantics intended by DCI…(see argument Joshua)

(so not being diff. semantics like in your example with and without hats)

 

So what you say is:

Make a new own hasPart being a subclass of dcterms:haPart and then add the fact that my hasPart is of type owl:objectProperty

To be even more cleaner and independent of my assumption that I actually model what was probably meant by DCI…

 

???

 

 

Dr. ir. H.M. (Michel) Bohms
Sr. Research Scientist
Structural Reliability

T +31 (0)88 866 31 07
M +31 (0)63 038 12 20
[hidden email]

Location

 

This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.

 

From: protege-user [[hidden email]] On Behalf Of Ron Rudnicki
Sent: woensdag 30 april 2014 15:23
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] FW: advice on adding info to other people's ontology

 

But consider. I take your class: Person and to it I add an axiom that Person is a subclass of things that wear hats. Now I publish my ontology claiming that the class person is the same as yours. I also collect instance data about Persons using my class. You try to aggregrate my data with yours, funny thing is that all these people wear hats…

 

What I should have done instead was take your class as a parent class and create a subclass called PersonWithHat or HattedPerson. Then there’s no problem.

 

From: protege-user [[hidden email]] On Behalf Of Lieke Verhelst
Sent: Wednesday, April 30, 2014 9:03 AM
To: [hidden email]
Subject: [protege-user] FW: advice on adding info to other people's ontology

 

 

Hi all,
 
Adding statements to an entity from a namespace that is not yours the way
Michel suggests is saying things about someone else's things like you
were that other party.  It is assuming that this is technically possible.
( It isn't, because you cannot publish the statement in someone else's
namespace since it is on someone else's server.) 
The editors allow it but imho this type of practice must be avoided
because it causes unexpected results, just like Ron says. I haven't
checked but hopefully reasoners act on the real origin of statements on
the web and ignore the statements about remote namespaces that are locally
declared.
 
Of course it is perfectly fine to add statements in your own namespace
that say something about someone else's entities.
 
 
Best, Lieke
Linked Data Factory
Semantic web and Linked Data Services
The Netherlands
 
-----Original Message-----
From: Bohms, H.M. (Michel) [mailto:[hidden email]] 
Sent: 29 April 2014 19:11
To: [hidden email]'; 'User support for WebProtege and Protege
Desktop'
Cc: 'Lieke Verhelst'
Subject: RE: [protege-user] advice on adding info to other people's
ontology
 
Great, so I actually use it in the spirit of.
I'll keep it the way I have it (with the triple added on my side)
 
Thx Joshua for your clear analysis, Michel
 
-----Original Message-----
From: protege-user [mailto:[hidden email]] On
Behalf Of Joshua TAYLOR
Sent: dinsdag 29 april 2014 17:19
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] advice on adding info to other people's
ontology
 
On Tue, Apr 29, 2014 at 11:15 AM, Joshua TAYLOR <[hidden email]>
wrote:

If you have a property op_hasPart that's a subproperty of some DC property, then the DC property will still need to be declared as either an annotation property, a datatype property, or an object property, and the only one that makes sense here would be as an object property. So this doesn't really avoid the problem.

it is worth noting though, that the documentation about hasPart [1]
says: "This term is intended to be used with non-literal values as defined
in the DCMI Abstract Model
(http://dublincore.org/documents/abstract-model/). As of December 2007,
the DCMI Usage Board is seeking a way to express this intention with a
formal range declaration."  It doesn't have a formal range, but the fact
that it's intended to be used with non-literal values at least means that
a declaration as an owl:ObjectProperty won't be too far from its intended
use.  If nothing else, the values it will have as an object property will
be non-literals, so any triples with the owl:ObjectProperty hasParts will
be compatible the rdf:Property hasPart.
 
//JT
 
[1] http://dublincore.org/documents/dcmi-terms/#terms-hasPart
 
--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
 
 
 
Dit bericht kan informatie bevatten die niet voor u is bestemd. Indien u
niet de geadresseerde bent of dit bericht abusievelijk aan u is
toegezonden, wordt u verzocht dat aan de afzender te melden en het bericht
te verwijderen. TNO aanvaardt geen aansprakelijkheid voor de inhoud van
deze e-mail, de wijze waarop u deze gebruikt en voor schade, van welke
aard ook, die verband houdt met risico's verbonden aan het elektronisch
verzenden van berichten.
 
 
 
This message may contain information that is not intended for you. If you
are not the addressee or if this message was sent to you by mistake, you
are requested to inform the sender and delete the message. TNO accepts no
liability for the content of this e-mail, for the manner in which you use
it and for damage of any kind resulting from the risks inherent to the
electronic transmission of messages.
 

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: FW: advice on adding info to other people's ontology

Timothy Redmond
In reply to this post by Ron Rudnicki
On 04/30/2014 06:23 AM, Ron Rudnicki wrote:

But consider. I take your class: Person and to it I add an axiom that Person is a subclass of things that wear hats. Now I publish my ontology claiming that the class person is the same as yours. I also collect instance data about Persons using my class. You try to aggregrate my data with yours, funny thing is that all these people wear hats…

 

What I should have done instead was take your class as a parent class and create a subclass called PersonWithHat or HattedPerson. Then there’s no problem.


As a general principle I don't agree with this.  I think that axioms in OWL ontologies are analogous to sentences in a book.  If I am writing a book I may add a statement like "Humans are spiritual beings."  If I am forced by my editor to rewrite this as "Spiritual being humans are spiritual beings" the meaning of my statement is changed.

This also gets into the issue of what is a definition and what is not.  When I make my statement about humans it may be important to me that I am not redefining the concept.  I am re-using someone else's concept (perhaps obtained from a book about physiology) and making a statement about that concept.  I think that in OWL it is also true that just because something might look like a definition, it does not mean that the statement is to be taken as a definition.  I feel that a better approach is to look at the axioms individually and see if you believe that they are "true" (in some sense).

Now the case of this discussion is a bit different because it deals with declarations.  If I wrote a book and declared that "human" is an adverb then you might quarrel with me and say that I am using the word in a non-standard manner.  Similarly you could argue, perhaps, that changing the declaration of a IRI is not a good thing to do. 

However, while I haven't checked today, I think that the dublin core ontology is a bit short on declarations which, if true, makes it unsuitable (as is) for OWL work.  So in that case, I think that giving it a declaration that you think is compatible with the definition of the vocabulary is arguably appropriate.

-Timothy


 

From: protege-user [[hidden email]] On Behalf Of Lieke Verhelst
Sent: Wednesday, April 30, 2014 9:03 AM
To: [hidden email]
Subject: [protege-user] FW: advice on adding info to other people's ontology

 

 

Hi all,
 
Adding statements to an entity from a namespace that is not yours the way
Michel suggests is saying things about someone else's things like you
were that other party.  It is assuming that this is technically possible.
( It isn't, because you cannot publish the statement in someone else's
namespace since it is on someone else's server.) 
The editors allow it but imho this type of practice must be avoided
because it causes unexpected results, just like Ron says. I haven't
checked but hopefully reasoners act on the real origin of statements on
the web and ignore the statements about remote namespaces that are locally
declared.
 
Of course it is perfectly fine to add statements in your own namespace
that say something about someone else's entities.
 
 
Best, Lieke
Linked Data Factory
Semantic web and Linked Data Services
The Netherlands
 
-----Original Message-----
From: Bohms, H.M. (Michel) [mailto:[hidden email]] 
Sent: 29 April 2014 19:11
To: [hidden email]'; 'User support for WebProtege and Protege
Desktop'
Cc: 'Lieke Verhelst'
Subject: RE: [protege-user] advice on adding info to other people's
ontology
 
Great, so I actually use it in the spirit of.
I'll keep it the way I have it (with the triple added on my side)
 
Thx Joshua for your clear analysis, Michel
 
-----Original Message-----
From: protege-user [mailto:[hidden email]] On
Behalf Of Joshua TAYLOR
Sent: dinsdag 29 april 2014 17:19
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] advice on adding info to other people's
ontology
 
On Tue, Apr 29, 2014 at 11:15 AM, Joshua TAYLOR <[hidden email]>
wrote:

If you have a property op_hasPart that's a subproperty of some DC property, then the DC property will still need to be declared as either an annotation property, a datatype property, or an object property, and the only one that makes sense here would be as an object property. So this doesn't really avoid the problem.

it is worth noting though, that the documentation about hasPart [1]
says: "This term is intended to be used with non-literal values as defined
in the DCMI Abstract Model
(http://dublincore.org/documents/abstract-model/). As of December 2007,
the DCMI Usage Board is seeking a way to express this intention with a
formal range declaration."  It doesn't have a formal range, but the fact
that it's intended to be used with non-literal values at least means that
a declaration as an owl:ObjectProperty won't be too far from its intended
use.  If nothing else, the values it will have as an object property will
be non-literals, so any triples with the owl:ObjectProperty hasParts will
be compatible the rdf:Property hasPart.
 
//JT
 
[1] http://dublincore.org/documents/dcmi-terms/#terms-hasPart
 
--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
 
 
 
Dit bericht kan informatie bevatten die niet voor u is bestemd. Indien u
niet de geadresseerde bent of dit bericht abusievelijk aan u is
toegezonden, wordt u verzocht dat aan de afzender te melden en het bericht
te verwijderen. TNO aanvaardt geen aansprakelijkheid voor de inhoud van
deze e-mail, de wijze waarop u deze gebruikt en voor schade, van welke
aard ook, die verband houdt met risico's verbonden aan het elektronisch
verzenden van berichten.
 
 
 
This message may contain information that is not intended for you. If you
are not the addressee or if this message was sent to you by mistake, you
are requested to inform the sender and delete the message. TNO accepts no
liability for the content of this e-mail, for the manner in which you use
it and for damage of any kind resulting from the risks inherent to the
electronic transmission of messages.
 


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


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: FW: advice on adding info to other people's ontology

Lieke Verhelst
In reply to this post by Ron Rudnicki
Hi Joshua,

can you please explain:

if you allow that anybody can publish additional axiom statements about a remote source in their own ontology, like we are discussing here:
example <http://purl.org/dc/terms#hasPart> rdf:type owl:ObjectProperty .

how can the original source maintain what it intended to say ?
What about dereferencability (or do you consider this a Linked Data issue?)

thanks, Lieke



From: [hidden email]
To: [hidden email]
Date: Wed, 30 Apr 2014 10:58:15 -0400
Subject: Re: [protege-user] FW: advice on adding info to other people's ontology

I think this is a good summary.

 

From: protege-user [mailto:[hidden email]] On Behalf Of Bohms, H.M. (Michel)
Sent: Wednesday, April 30, 2014 10:30 AM
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] FW: advice on adding info to other people's ontology

 

Can we summarise (?):

1.       Best way: make 100% sure no issues: subclass own from source or reinvent (without subclassing)

2.       Reasonable way: assume something that is somehow missing in the authorative source which is not contradicted by the source

a.       What would be IN THIS CASE the “unexpected result” by a reasoner?

3.       Do not: assume things that likely contradict with explicit/implicit statements by the source (hats example)

 

This whole discussion does not only aply to dcterms:hasPart but also to the qudt units we reuse.

I assume they also fall in category 2. Since the sources describes several use cases where for each use case you might have to add some stuff yourself first to make it work

(here: add the triple: <http://qudt.org/schema/qudt#Unit> rdfs:subClassOf rdfs:Datatype .)

 

 

 

 

 

 

Dr. ir. H.M. (Michel) Bohms
Sr. Research Scientist
Structural Reliability

T +31 (0)88 866 31 07
M +31 (0)63 038 12 20
[hidden email]

Location

 

This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.

 

From: protege-user [[hidden email]] On Behalf Of Ron Rudnicki
Sent: woensdag 30 april 2014 16:05
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] FW: advice on adding info to other people's ontology

 

If my caution applies at all, it applies only if DC publishes an .owl file in which dcterms:hasPart is typed as something other than owl:ObjectProperty. If it does, then by changing  the type in your file there could be queries that unexpectedly return results from your ontology but not from another that uses the canonical DC file or vice versa. In this case if you subtype, then you are warning your users that some extra effort will be needed to merge data, thereby saving us some debugging efforts. Of course, If DC offers no guidance on the implementation of hasPart, then typing it for one’s own namespace is both legitimate and necessary.

 

From: protege-user [[hidden email]] On Behalf Of Bohms, H.M. (Michel)
Sent: Wednesday, April 30, 2014 9:32 AM
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] FW: advice on adding info to other people's ontology

 

I assumed it works in MY/THIS case because I added a triple that was actually in consideration of / in the spirit of the semantics intended by DCI…(see argument Joshua)

(so not being diff. semantics like in your example with and without hats)

 

So what you say is:

Make a new own hasPart being a subclass of dcterms:haPart and then add the fact that my hasPart is of type owl:objectProperty

To be even more cleaner and independent of my assumption that I actually model what was probably meant by DCI…

 

???

 

 

Dr. ir. H.M. (Michel) Bohms
Sr. Research Scientist
Structural Reliability

T +31 (0)88 866 31 07
M +31 (0)63 038 12 20
[hidden email]

Location

 

This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.

 

From: protege-user [[hidden email]] On Behalf Of Ron Rudnicki
Sent: woensdag 30 april 2014 15:23
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] FW: advice on adding info to other people's ontology

 

But consider. I take your class: Person and to it I add an axiom that Person is a subclass of things that wear hats. Now I publish my ontology claiming that the class person is the same as yours. I also collect instance data about Persons using my class. You try to aggregrate my data with yours, funny thing is that all these people wear hats…

 

What I should have done instead was take your class as a parent class and create a subclass called PersonWithHat or HattedPerson. Then there’s no problem.

 

From: protege-user [[hidden email]] On Behalf Of Lieke Verhelst
Sent: Wednesday, April 30, 2014 9:03 AM
To: [hidden email]
Subject: [protege-user] FW: advice on adding info to other people's ontology

 

 

Hi all,
 
Adding statements to an entity from a namespace that is not yours the way
Michel suggests is saying things about someone else's things like you
were that other party.  It is assuming that this is technically possible.
( It isn't, because you cannot publish the statement in someone else's
namespace since it is on someone else's server.) 
The editors allow it but imho this type of practice must be avoided
because it causes unexpected results, just like Ron says. I haven't
checked but hopefully reasoners act on the real origin of statements on
the web and ignore the statements about remote namespaces that are locally
declared.
 
Of course it is perfectly fine to add statements in your own namespace
that say something about someone else's entities.
 
 
Best, Lieke
Linked Data Factory
Semantic web and Linked Data Services
The Netherlands
 
-----Original Message-----
From: Bohms, H.M. (Michel) [mailto:[hidden email]] 
Sent: 29 April 2014 19:11
To: [hidden email]'; 'User support for WebProtege and Protege
Desktop'
Cc: 'Lieke Verhelst'
Subject: RE: [protege-user] advice on adding info to other people's
ontology
 
Great, so I actually use it in the spirit of.
I'll keep it the way I have it (with the triple added on my side)
 
Thx Joshua for your clear analysis, Michel
 
-----Original Message-----
From: protege-user [mailto:[hidden email]] On
Behalf Of Joshua TAYLOR
Sent: dinsdag 29 april 2014 17:19
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] advice on adding info to other people's
ontology
 
On Tue, Apr 29, 2014 at 11:15 AM, Joshua TAYLOR <[hidden email]>
wrote:

If you have a property op_hasPart that's a subproperty of some DC property, then the DC property will still need to be declared as either an annotation property, a datatype property, or an object property, and the only one that makes sense here would be as an object property. So this doesn't really avoid the problem.

it is worth noting though, that the documentation about hasPart [1]
says: "This term is intended to be used with non-literal values as defined
in the DCMI Abstract Model
(http://dublincore.org/documents/abstract-model/). As of December 2007,
the DCMI Usage Board is seeking a way to express this intention with a
formal range declaration."  It doesn't have a formal range, but the fact
that it's intended to be used with non-literal values at least means that
a declaration as an owl:ObjectProperty won't be too far from its intended
use.  If nothing else, the values it will have as an object property will
be non-literals, so any triples with the owl:ObjectProperty hasParts will
be compatible the rdf:Property hasPart.
 
//JT
 
[1] http://dublincore.org/documents/dcmi-terms/#terms-hasPart
 
--
Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
 
 
 
Dit bericht kan informatie bevatten die niet voor u is bestemd. Indien u
niet de geadresseerde bent of dit bericht abusievelijk aan u is
toegezonden, wordt u verzocht dat aan de afzender te melden en het bericht
te verwijderen. TNO aanvaardt geen aansprakelijkheid voor de inhoud van
deze e-mail, de wijze waarop u deze gebruikt en voor schade, van welke
aard ook, die verband houdt met risico's verbonden aan het elektronisch
verzenden van berichten.
 
 
 
This message may contain information that is not intended for you. If you
are not the addressee or if this message was sent to you by mistake, you
are requested to inform the sender and delete the message. TNO accepts no
liability for the content of this e-mail, for the manner in which you use
it and for damage of any kind resulting from the risks inherent to the
electronic transmission of messages.
 

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

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: FW: advice on adding info to other people's ontology

Eric Prud'hommeaux
In reply to this post by Bohms, H.M. (Michel)
dcterms:hasPart is typed as something other than owl:ObjectProperty

* Bohms, H.M. (Michel) <[hidden email]> [2014-04-30 14:29+0000]
> Can we summarise (?):
>
> 1.      Best way: make 100% sure no issues: subclass own from source or reinvent (without subclassing)
>
> 2.      Reasonable way: assume something that is somehow missing in the authorative source which is not contradicted by the source
>
> a.      What would be IN THIS CASE the "unexpected result" by a reasoner?
>
> 3.      Do not: assume things that likely contradict with explicit/implicit statements by the source (hats example)

I think the only "safe" rule above is 1. That is, if someone accepts your verson of an ontology as a replacement for the authoritative copy, any changes to the original ontology can result in contradictions or overconstraints.

2 differs from 3 only in degree. If someone out there has data using dcterms:hasPart or dcterms:author as a DatatypeProperty, the modified ontology no longer fits their data, just as if the modified ontology were to say that all people where hats.

It's really practical to create axioms which further constrain one's own data and people will be tempted to do this by patching ontologies. They have to evaluate whether it can cause unexpected consequneces with unsuspecting third parties as well as whether having a skewed reality will affect their ability to consume outside data.


> This whole discussion does not only aply to dcterms:hasPart but also to the qudt units we reuse.
> I assume they also fall in category 2. Since the sources describes several use cases where for each use case you might have to add some stuff yourself first to make it work
> (here: add the triple: <http://qudt.org/schema/qudt#Unit> rdfs:subClassOf rdfs:Datatype .)
>
>
> Dr. ir. H.M. (Michel) Bohms
> Sr. Research Scientist
> Structural Reliability
>
> T +31 (0)88 866 31 07
> M +31 (0)63 038 12 20
> E [hidden email]<mailto:[hidden email]>
>
> Location<http://www.tno.nl/locaties/DTM>
>
>
>
> [cid:image001.gif@01CF6491.1F902690]<http://www.tno.nl/>
>
> This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.
>
>
>
> From: protege-user [mailto:[hidden email]] On Behalf Of Ron Rudnicki
> Sent: woensdag 30 april 2014 16:05
> To: User support for WebProtege and Protege Desktop
> Subject: Re: [protege-user] FW: advice on adding info to other people's ontology
>
> If my caution applies at all, it applies only if DC publishes an .owl file in which dcterms:hasPart is typed as something other than owl:ObjectProperty. If it does, then by changing  the type in your file there could be queries that unexpectedly return results from your ontology but not from another that uses the canonical DC file or vice versa. In this case if you subtype, then you are warning your users that some extra effort will be needed to merge data, thereby saving us some debugging efforts. Of course, If DC offers no guidance on the implementation of hasPart, then typing it for one's own namespace is both legitimate and necessary.
>
> From: protege-user [mailto:[hidden email]] On Behalf Of Bohms, H.M. (Michel)
> Sent: Wednesday, April 30, 2014 9:32 AM
> To: User support for WebProtege and Protege Desktop
> Subject: Re: [protege-user] FW: advice on adding info to other people's ontology
>
> I assumed it works in MY/THIS case because I added a triple that was actually in consideration of / in the spirit of the semantics intended by DCI...(see argument Joshua)
> (so not being diff. semantics like in your example with and without hats)
>
> So what you say is:
> Make a new own hasPart being a subclass of dcterms:haPart and then add the fact that my hasPart is of type owl:objectProperty
> To be even more cleaner and independent of my assumption that I actually model what was probably meant by DCI...
>
> ???
>
>
>
> Dr. ir. H.M. (Michel) Bohms
> Sr. Research Scientist
> Structural Reliability
>
> T +31 (0)88 866 31 07
> M +31 (0)63 038 12 20
> E [hidden email]<mailto:[hidden email]>
>
> Location<http://www.tno.nl/locaties/DTM>
>
>
>
> [cid:image001.gif@01CF6491.1F902690]<http://www.tno.nl/>
>
> This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.
>
>
>
> From: protege-user [mailto:[hidden email]] On Behalf Of Ron Rudnicki
> Sent: woensdag 30 april 2014 15:23
> To: User support for WebProtege and Protege Desktop
> Subject: Re: [protege-user] FW: advice on adding info to other people's ontology
>
> But consider. I take your class: Person and to it I add an axiom that Person is a subclass of things that wear hats. Now I publish my ontology claiming that the class person is the same as yours. I also collect instance data about Persons using my class. You try to aggregrate my data with yours, funny thing is that all these people wear hats...
>
> What I should have done instead was take your class as a parent class and create a subclass called PersonWithHat or HattedPerson. Then there's no problem.
>
> From: protege-user [mailto:[hidden email]] On Behalf Of Lieke Verhelst
> Sent: Wednesday, April 30, 2014 9:03 AM
> To: [hidden email]<mailto:[hidden email]>
> Subject: [protege-user] FW: advice on adding info to other people's ontology
>
>
>
> Hi all,
>
>
>
> Adding statements to an entity from a namespace that is not yours the way
>
> Michel suggests is saying things about someone else's things like you
>
> were that other party.  It is assuming that this is technically possible.
>
> ( It isn't, because you cannot publish the statement in someone else's
>
> namespace since it is on someone else's server.)
>
> The editors allow it but imho this type of practice must be avoided
>
> because it causes unexpected results, just like Ron says. I haven't
>
> checked but hopefully reasoners act on the real origin of statements on
>
> the web and ignore the statements about remote namespaces that are locally
>
> declared.
>
>
>
> Of course it is perfectly fine to add statements in your own namespace
>
> that say something about someone else's entities.
>
>
>
>
>
> Best, Lieke
>
> Linked Data Factory
>
> Semantic web and Linked Data Services
>
> The Netherlands
>
>
>
> -----Original Message-----
>
> From: Bohms, H.M. (Michel) [mailto:[hidden email]<mailto:[hidden email]>]
>
> Sent: 29 April 2014 19:11
>
> To: '[hidden email]<mailto:'[hidden email]>'; 'User support for WebProtege and Protege
>
> Desktop'
>
> Cc: 'Lieke Verhelst'
>
> Subject: RE: [protege-user] advice on adding info to other people's
>
> ontology
>
>
>
> Great, so I actually use it in the spirit of.
>
> I'll keep it the way I have it (with the triple added on my side)
>
>
>
> Thx Joshua for your clear analysis, Michel
>
>
>
> -----Original Message-----
>
> From: protege-user [mailto:[hidden email]<mailto:[hidden email]>] On
>
> Behalf Of Joshua TAYLOR
>
> Sent: dinsdag 29 april 2014 17:19
>
> To: User support for WebProtege and Protege Desktop
>
> Subject: Re: [protege-user] advice on adding info to other people's
>
> ontology
>
>
>
> On Tue, Apr 29, 2014 at 11:15 AM, Joshua TAYLOR <[hidden email]<mailto:[hidden email]>>
>
> wrote:
> If you have a property op_hasPart that's a subproperty of some DC property, then the DC property will still need to be declared as either an annotation property, a datatype property, or an object property, and the only one that makes sense here would be as an object property. So this doesn't really avoid the problem.
>
> it is worth noting though, that the documentation about hasPart [1]
>
> says: "This term is intended to be used with non-literal values as defined
>
> in the DCMI Abstract Model
>
> (http://dublincore.org/documents/abstract-model/). As of December 2007,
>
> the DCMI Usage Board is seeking a way to express this intention with a
>
> formal range declaration."  It doesn't have a formal range, but the fact
>
> that it's intended to be used with non-literal values at least means that
>
> a declaration as an owl:ObjectProperty won't be too far from its intended
>
> use.  If nothing else, the values it will have as an object property will
>
> be non-literals, so any triples with the owl:ObjectProperty hasParts will
>
> be compatible the rdf:Property hasPart.
>
>
>
> //JT
>
>
>
> [1] http://dublincore.org/documents/dcmi-terms/#terms-hasPart
>
>
>
> --
>
> Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
>
> _______________________________________________
>
> protege-user mailing list
>
> [hidden email]<mailto:[hidden email]>
>
> https://mailman.stanford.edu/mailman/listinfo/protege-user
>
>
>
>
>
>
>
> Dit bericht kan informatie bevatten die niet voor u is bestemd. Indien u
>
> niet de geadresseerde bent of dit bericht abusievelijk aan u is
>
> toegezonden, wordt u verzocht dat aan de afzender te melden en het bericht
>
> te verwijderen. TNO aanvaardt geen aansprakelijkheid voor de inhoud van
>
> deze e-mail, de wijze waarop u deze gebruikt en voor schade, van welke
>
> aard ook, die verband houdt met risico's verbonden aan het elektronisch
>
> verzenden van berichten.
>
>
>
>
>
>
>
> This message may contain information that is not intended for you. If you
>
> are not the addressee or if this message was sent to you by mistake, you
>
> are requested to inform the sender and delete the message. TNO accepts no
>
> liability for the content of this e-mail, for the manner in which you use
>
> it and for damage of any kind resulting from the risks inherent to the
>
> electronic transmission of messages.
>
>



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


--
-ericP

office: +1.617.599.3509
mobile: +33.6.80.80.35.59

([hidden email])
Feel free to forward this message to any list for any purpose other than
email address distribution.

There are subtle nuances encoded in font variation and clever layout
which can only be seen by printing this message on high-clay paper.
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: FW: advice on adding info to other people's ontology

Bohms, H.M. (Michel)
Hi Eric,

Can you claify you point? (The whole issue is about difernt rdf:type's by different people)

Thx Michel

-----Original Message-----
From: protege-user [mailto:[hidden email]] On Behalf Of Eric Prud'hommeaux
Sent: donderdag 1 mei 2014 6:54
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] FW: advice on adding info to other people's ontology

dcterms:hasPart is typed as something other than owl:ObjectProperty

* Bohms, H.M. (Michel) <[hidden email]> [2014-04-30 14:29+0000]
> Can we summarise (?):
>
> 1.      Best way: make 100% sure no issues: subclass own from source or reinvent (without subclassing)
>
> 2.      Reasonable way: assume something that is somehow missing in the authorative source which is not contradicted by the source
>
> a.      What would be IN THIS CASE the "unexpected result" by a reasoner?
>
> 3.      Do not: assume things that likely contradict with explicit/implicit statements by the source (hats example)

I think the only "safe" rule above is 1. That is, if someone accepts your verson of an ontology as a replacement for the authoritative copy, any changes to the original ontology can result in contradictions or overconstraints.

2 differs from 3 only in degree. If someone out there has data using dcterms:hasPart or dcterms:author as a DatatypeProperty, the modified ontology no longer fits their data, just as if the modified ontology were to say that all people where hats.

It's really practical to create axioms which further constrain one's own data and people will be tempted to do this by patching ontologies. They have to evaluate whether it can cause unexpected consequneces with unsuspecting third parties as well as whether having a skewed reality will affect their ability to consume outside data.


> This whole discussion does not only aply to dcterms:hasPart but also to the qudt units we reuse.
> I assume they also fall in category 2. Since the sources describes
> several use cases where for each use case you might have to add some
> stuff yourself first to make it work
> (here: add the triple: <http://qudt.org/schema/qudt#Unit>
> rdfs:subClassOf rdfs:Datatype .)
>
>
> Dr. ir. H.M. (Michel) Bohms
> Sr. Research Scientist
> Structural Reliability
>
> T +31 (0)88 866 31 07
> M +31 (0)63 038 12 20
> E [hidden email]<mailto:[hidden email]>
>
> Location<http://www.tno.nl/locaties/DTM>
>
>
>
> [cid:image001.gif@01CF6491.1F902690]<http://www.tno.nl/>
>
> This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.
>
>
>
> From: protege-user [mailto:[hidden email]] On
> Behalf Of Ron Rudnicki
> Sent: woensdag 30 april 2014 16:05
> To: User support for WebProtege and Protege Desktop
> Subject: Re: [protege-user] FW: advice on adding info to other
> people's ontology
>
> If my caution applies at all, it applies only if DC publishes an .owl file in which dcterms:hasPart is typed as something other than owl:ObjectProperty. If it does, then by changing  the type in your file there could be queries that unexpectedly return results from your ontology but not from another that uses the canonical DC file or vice versa. In this case if you subtype, then you are warning your users that some extra effort will be needed to merge data, thereby saving us some debugging efforts. Of course, If DC offers no guidance on the implementation of hasPart, then typing it for one's own namespace is both legitimate and necessary.
>
> From: protege-user [mailto:[hidden email]] On
> Behalf Of Bohms, H.M. (Michel)
> Sent: Wednesday, April 30, 2014 9:32 AM
> To: User support for WebProtege and Protege Desktop
> Subject: Re: [protege-user] FW: advice on adding info to other
> people's ontology
>
> I assumed it works in MY/THIS case because I added a triple that was
> actually in consideration of / in the spirit of the semantics intended
> by DCI...(see argument Joshua) (so not being diff. semantics like in
> your example with and without hats)
>
> So what you say is:
> Make a new own hasPart being a subclass of dcterms:haPart and then add
> the fact that my hasPart is of type owl:objectProperty To be even more cleaner and independent of my assumption that I actually model what was probably meant by DCI...
>
> ???
>
>
>
> Dr. ir. H.M. (Michel) Bohms
> Sr. Research Scientist
> Structural Reliability
>
> T +31 (0)88 866 31 07
> M +31 (0)63 038 12 20
> E [hidden email]<mailto:[hidden email]>
>
> Location<http://www.tno.nl/locaties/DTM>
>
>
>
> [cid:image001.gif@01CF6491.1F902690]<http://www.tno.nl/>
>
> This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.
>
>
>
> From: protege-user [mailto:[hidden email]] On
> Behalf Of Ron Rudnicki
> Sent: woensdag 30 april 2014 15:23
> To: User support for WebProtege and Protege Desktop
> Subject: Re: [protege-user] FW: advice on adding info to other
> people's ontology
>
> But consider. I take your class: Person and to it I add an axiom that Person is a subclass of things that wear hats. Now I publish my ontology claiming that the class person is the same as yours. I also collect instance data about Persons using my class. You try to aggregrate my data with yours, funny thing is that all these people wear hats...
>
> What I should have done instead was take your class as a parent class and create a subclass called PersonWithHat or HattedPerson. Then there's no problem.
>
> From: protege-user [mailto:[hidden email]] On
> Behalf Of Lieke Verhelst
> Sent: Wednesday, April 30, 2014 9:03 AM
> To:
> [hidden email]<mailto:[hidden email]
> >
> Subject: [protege-user] FW: advice on adding info to other people's
> ontology
>
>
>
> Hi all,
>
>
>
> Adding statements to an entity from a namespace that is not yours the
> way
>
> Michel suggests is saying things about someone else's things like you
>
> were that other party.  It is assuming that this is technically possible.
>
> ( It isn't, because you cannot publish the statement in someone else's
>
> namespace since it is on someone else's server.)
>
> The editors allow it but imho this type of practice must be avoided
>
> because it causes unexpected results, just like Ron says. I haven't
>
> checked but hopefully reasoners act on the real origin of statements
> on
>
> the web and ignore the statements about remote namespaces that are
> locally
>
> declared.
>
>
>
> Of course it is perfectly fine to add statements in your own namespace
>
> that say something about someone else's entities.
>
>
>
>
>
> Best, Lieke
>
> Linked Data Factory
>
> Semantic web and Linked Data Services
>
> The Netherlands
>
>
>
> -----Original Message-----
>
> From: Bohms, H.M. (Michel)
> [mailto:[hidden email]<mailto:[hidden email]>]
>
> Sent: 29 April 2014 19:11
>
> To: '[hidden email]<mailto:'[hidden email]>'; 'User
> support for WebProtege and Protege
>
> Desktop'
>
> Cc: 'Lieke Verhelst'
>
> Subject: RE: [protege-user] advice on adding info to other people's
>
> ontology
>
>
>
> Great, so I actually use it in the spirit of.
>
> I'll keep it the way I have it (with the triple added on my side)
>
>
>
> Thx Joshua for your clear analysis, Michel
>
>
>
> -----Original Message-----
>
> From: protege-user
> [mailto:[hidden email]<mailto:protege-user-bo
> [hidden email]>] On
>
> Behalf Of Joshua TAYLOR
>
> Sent: dinsdag 29 april 2014 17:19
>
> To: User support for WebProtege and Protege Desktop
>
> Subject: Re: [protege-user] advice on adding info to other people's
>
> ontology
>
>
>
> On Tue, Apr 29, 2014 at 11:15 AM, Joshua TAYLOR
> <[hidden email]<mailto:[hidden email]>>
>
> wrote:
> If you have a property op_hasPart that's a subproperty of some DC property, then the DC property will still need to be declared as either an annotation property, a datatype property, or an object property, and the only one that makes sense here would be as an object property. So this doesn't really avoid the problem.
>
> it is worth noting though, that the documentation about hasPart [1]
>
> says: "This term is intended to be used with non-literal values as
> defined
>
> in the DCMI Abstract Model
>
> (http://dublincore.org/documents/abstract-model/). As of December
> 2007,
>
> the DCMI Usage Board is seeking a way to express this intention with a
>
> formal range declaration."  It doesn't have a formal range, but the
> fact
>
> that it's intended to be used with non-literal values at least means
> that
>
> a declaration as an owl:ObjectProperty won't be too far from its
> intended
>
> use.  If nothing else, the values it will have as an object property
> will
>
> be non-literals, so any triples with the owl:ObjectProperty hasParts
> will
>
> be compatible the rdf:Property hasPart.
>
>
>
> //JT
>
>
>
> [1] http://dublincore.org/documents/dcmi-terms/#terms-hasPart
>
>
>
> --
>
> Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
>
> _______________________________________________
>
> protege-user mailing list
>
> [hidden email]<mailto:[hidden email]
> >
>
> https://mailman.stanford.edu/mailman/listinfo/protege-user
>
>
>
>
>
>
>
> Dit bericht kan informatie bevatten die niet voor u is bestemd. Indien
> u
>
> niet de geadresseerde bent of dit bericht abusievelijk aan u is
>
> toegezonden, wordt u verzocht dat aan de afzender te melden en het
> bericht
>
> te verwijderen. TNO aanvaardt geen aansprakelijkheid voor de inhoud
> van
>
> deze e-mail, de wijze waarop u deze gebruikt en voor schade, van welke
>
> aard ook, die verband houdt met risico's verbonden aan het
> elektronisch
>
> verzenden van berichten.
>
>
>
>
>
>
>
> This message may contain information that is not intended for you. If
> you
>
> are not the addressee or if this message was sent to you by mistake,
> you
>
> are requested to inform the sender and delete the message. TNO accepts
> no
>
> liability for the content of this e-mail, for the manner in which you
> use
>
> it and for damage of any kind resulting from the risks inherent to the
>
> electronic transmission of messages.
>
>



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


--
-ericP

office: +1.617.599.3509
mobile: +33.6.80.80.35.59

([hidden email])
Feel free to forward this message to any list for any purpose other than email address distribution.

There are subtle nuances encoded in font variation and clever layout which can only be seen by printing this message on high-clay paper.
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: FW: advice on adding info to other people's ontology

Lieke Verhelst
Hi all,
Answering my own question to Joshua: provenance to be checked at e.g. the graph level.
Thanks for all your feedback..interesting discussion..
best, Lieke

> From: [hidden email]

> To: [hidden email]
> Date: Thu, 1 May 2014 06:35:22 +0000
> Subject: Re: [protege-user] FW: advice on adding info to other people's ontology
>
> Hi Eric,
>
> Can you claify you point? (The whole issue is about difernt rdf:type's by different people)
>
> Thx Michel
>
> -----Original Message-----
> From: protege-user [mailto:[hidden email]] On Behalf Of Eric Prud'hommeaux
> Sent: donderdag 1 mei 2014 6:54
> To: User support for WebProtege and Protege Desktop
> Subject: Re: [protege-user] FW: advice on adding info to other people's ontology
>
> dcterms:hasPart is typed as something other than owl:ObjectProperty
>
> * Bohms, H.M. (Michel) <[hidden email]> [2014-04-30 14:29+0000]
> > Can we summarise (?):
> >
> > 1. Best way: make 100% sure no issues: subclass own from source or reinvent (without subclassing)
> >
> > 2. Reasonable way: assume something that is somehow missing in the authorative source which is not contradicted by the source
> >
> > a. What would be IN THIS CASE the "unexpected result" by a reasoner?
> >
> > 3. Do not: assume things that likely contradict with explicit/implicit statements by the source (hats example)
>
> I think the only "safe" rule above is 1. That is, if someone accepts your verson of an ontology as a replacement for the authoritative copy, any changes to the original ontology can result in contradictions or overconstraints.
>
> 2 differs from 3 only in degree. If someone out there has data using dcterms:hasPart or dcterms:author as a DatatypeProperty, the modified ontology no longer fits their data, just as if the modified ontology were to say that all people where hats.
>
> It's really practical to create axioms which further constrain one's own data and people will be tempted to do this by patching ontologies. They have to evaluate whether it can cause unexpected consequneces with unsuspecting third parties as well as whether having a skewed reality will affect their ability to consume outside data.
>
>
> > This whole discussion does not only aply to dcterms:hasPart but also to the qudt units we reuse.
> > I assume they also fall in category 2. Since the sources describes
> > several use cases where for each use case you might have to add some
> > stuff yourself first to make it work
> > (here: add the triple: <http://qudt.org/schema/qudt#Unit>
> > rdfs:subClassOf rdfs:Datatype .)
> >
> >
> > Dr. ir. H.M. (Michel) Bohms
> > Sr. Research Scientist
> > Structural Reliability
> >
> > T +31 (0)88 866 31 07
> > M +31 (0)63 038 12 20
> > E [hidden email]<mailto:[hidden email]>
> >
> > Location<http://www.tno.nl/locaties/DTM>
> >
> >
> >
> > [cid:image001.gif@01CF6491.1F902690]<http://www.tno.nl/>
> >
> > This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.
> >
> >
> >
> > From: protege-user [mailto:[hidden email]] On
> > Behalf Of Ron Rudnicki
> > Sent: woensdag 30 april 2014 16:05
> > To: User support for WebProtege and Protege Desktop
> > Subject: Re: [protege-user] FW: advice on adding info to other
> > people's ontology
> >
> > If my caution applies at all, it applies only if DC publishes an .owl file in which dcterms:hasPart is typed as something other than owl:ObjectProperty. If it does, then by changing the type in your file there could be queries that unexpectedly return results from your ontology but not from another that uses the canonical DC file or vice versa. In this case if you subtype, then you are warning your users that some extra effort will be needed to merge data, thereby saving us some debugging efforts. Of course, If DC offers no guidance on the implementation of hasPart, then typing it for one's own namespace is both legitimate and necessary.
> >
> > From: protege-user [mailto:[hidden email]] On
> > Behalf Of Bohms, H.M. (Michel)
> > Sent: Wednesday, April 30, 2014 9:32 AM
> > To: User support for WebProtege and Protege Desktop
> > Subject: Re: [protege-user] FW: advice on adding info to other
> > people's ontology
> >
> > I assumed it works in MY/THIS case because I added a triple that was
> > actually in consideration of / in the spirit of the semantics intended
> > by DCI...(see argument Joshua) (so not being diff. semantics like in
> > your example with and without hats)
> >
> > So what you say is:
> > Make a new own hasPart being a subclass of dcterms:haPart and then add
> > the fact that my hasPart is of type owl:objectProperty To be even more cleaner and independent of my assumption that I actually model what was probably meant by DCI...
> >
> > ???
> >
> >
> >
> > Dr. ir. H.M. (Michel) Bohms
> > Sr. Research Scientist
> > Structural Reliability
> >
> > T +31 (0)88 866 31 07
> > M +31 (0)63 038 12 20
> > E [hidden email]<mailto:[hidden email]>
> >
> > Location<http://www.tno.nl/locaties/DTM>
> >
> >
> >
> > [cid:image001.gif@01CF6491.1F902690]<http://www.tno.nl/>
> >
> > This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.
> >
> >
> >
> > From: protege-user [mailto:[hidden email]] On
> > Behalf Of Ron Rudnicki
> > Sent: woensdag 30 april 2014 15:23
> > To: User support for WebProtege and Protege Desktop
> > Subject: Re: [protege-user] FW: advice on adding info to other
> > people's ontology
> >
> > But consider. I take your class: Person and to it I add an axiom that Person is a subclass of things that wear hats. Now I publish my ontology claiming that the class person is the same as yours. I also collect instance data about Persons using my class. You try to aggregrate my data with yours, funny thing is that all these people wear hats...
> >
> > What I should have done instead was take your class as a parent class and create a subclass called PersonWithHat or HattedPerson. Then there's no problem.
> >
> > From: protege-user [mailto:[hidden email]] On
> > Behalf Of Lieke Verhelst
> > Sent: Wednesday, April 30, 2014 9:03 AM
> > To:
> > [hidden email]<mailto:[hidden email]
> > >
> > Subject: [protege-user] FW: advice on adding info to other people's
> > ontology
> >
> >
> >
> > Hi all,
> >
> >
> >
> > Adding statements to an entity from a namespace that is not yours the
> > way
> >
> > Michel suggests is saying things about someone else's things like you
> >
> > were that other party. It is assuming that this is technically possible.
> >
> > ( It isn't, because you cannot publish the statement in someone else's
> >
> > namespace since it is on someone else's server.)
> >
> > The editors allow it but imho this type of practice must be avoided
> >
> > because it causes unexpected results, just like Ron says. I haven't
> >
> > checked but hopefully reasoners act on the real origin of statements
> > on
> >
> > the web and ignore the statements about remote namespaces that are
> > locally
> >
> > declared.
> >
> >
> >
> > Of course it is perfectly fine to add statements in your own namespace
> >
> > that say something about someone else's entities.
> >
> >
> >
> >
> >
> > Best, Lieke
> >
> > Linked Data Factory
> >
> > Semantic web and Linked Data Services
> >
> > The Netherlands
> >
> >
> >
> > -----Original Message-----
> >
> > From: Bohms, H.M. (Michel)
> > [mailto:[hidden email]<mailto:[hidden email]>]
> >
> > Sent: 29 April 2014 19:11
> >
> > To: '[hidden email]<mailto:'[hidden email]>'; 'User
> > support for WebProtege and Protege
> >
> > Desktop'
> >
> > Cc: 'Lieke Verhelst'
> >
> > Subject: RE: [protege-user] advice on adding info to other people's
> >
> > ontology
> >
> >
> >
> > Great, so I actually use it in the spirit of.
> >
> > I'll keep it the way I have it (with the triple added on my side)
> >
> >
> >
> > Thx Joshua for your clear analysis, Michel
> >
> >
> >
> > -----Original Message-----
> >
> > From: protege-user
> > [mailto:[hidden email]<mailto:protege-user-bo
> > [hidden email]>] On
> >
> > Behalf Of Joshua TAYLOR
> >
> > Sent: dinsdag 29 april 2014 17:19
> >
> > To: User support for WebProtege and Protege Desktop
> >
> > Subject: Re: [protege-user] advice on adding info to other people's
> >
> > ontology
> >
> >
> >
> > On Tue, Apr 29, 2014 at 11:15 AM, Joshua TAYLOR
> > <[hidden email]<mailto:[hidden email]>>
> >
> > wrote:
> > If you have a property op_hasPart that's a subproperty of some DC property, then the DC property will still need to be declared as either an annotation property, a datatype property, or an object property, and the only one that makes sense here would be as an object property. So this doesn't really avoid the problem.
> >
> > it is worth noting though, that the documentation about hasPart [1]
> >
> > says: "This term is intended to be used with non-literal values as
> > defined
> >
> > in the DCMI Abstract Model
> >
> > (http://dublincore.org/documents/abstract-model/). As of December
> > 2007,
> >
> > the DCMI Usage Board is seeking a way to express this intention with a
> >
> > formal range declaration." It doesn't have a formal range, but the
> > fact
> >
> > that it's intended to be used with non-literal values at least means
> > that
> >
> > a declaration as an owl:ObjectProperty won't be too far from its
> > intended
> >
> > use. If nothing else, the values it will have as an object property
> > will
> >
> > be non-literals, so any triples with the owl:ObjectProperty hasParts
> > will
> >
> > be compatible the rdf:Property hasPart.
> >
> >
> >
> > //JT
> >
> >
> >
> > [1] http://dublincore.org/documents/dcmi-terms/#terms-hasPart
> >
> >
> >
> > --
> >
> > Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
> >
> > _______________________________________________
> >
> > protege-user mailing list
> >
> > [hidden email]<mailto:[hidden email]
> > >
> >
> > https://mailman.stanford.edu/mailman/listinfo/protege-user
> >
> >
> >
> >
> >
> >
> >
> > Dit bericht kan informatie bevatten die niet voor u is bestemd. Indien
> > u
> >
> > niet de geadresseerde bent of dit bericht abusievelijk aan u is
> >
> > toegezonden, wordt u verzocht dat aan de afzender te melden en het
> > bericht
> >
> > te verwijderen. TNO aanvaardt geen aansprakelijkheid voor de inhoud
> > van
> >
> > deze e-mail, de wijze waarop u deze gebruikt en voor schade, van welke
> >
> > aard ook, die verband houdt met risico's verbonden aan het
> > elektronisch
> >
> > verzenden van berichten.
> >
> >
> >
> >
> >
> >
> >
> > This message may contain information that is not intended for you. If
> > you
> >
> > are not the addressee or if this message was sent to you by mistake,
> > you
> >
> > are requested to inform the sender and delete the message. TNO accepts
> > no
> >
> > liability for the content of this e-mail, for the manner in which you
> > use
> >
> > it and for damage of any kind resulting from the risks inherent to the
> >
> > electronic transmission of messages.
> >
> >
>
>
>
> > _______________________________________________
> > protege-user mailing list
> > [hidden email]
> > https://mailman.stanford.edu/mailman/listinfo/protege-user
>
>
> --
> -ericP
>
> office: +1.617.599.3509
> mobile: +33.6.80.80.35.59
>
> ([hidden email])
> Feel free to forward this message to any list for any purpose other than email address distribution.
>
> There are subtle nuances encoded in font variation and clever layout which can only be seen by printing this message on high-clay paper.
> _______________________________________________
> protege-user mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-user
> _______________________________________________
> protege-user mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-user

_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: FW: advice on adding info to other people's ontology

Bohms, H.M. (Michel)

Indeed interesting, I learned a lot, thx to all

 

Does anybody know where THIS kind of info is actually described? (in what kind of OWL document or specific papers/books)

 

Its so essential but somehow so difficult to trace … (guess its somewhere in the theoretic owl docs that shocked me too much too read….)

 

 

Dr. ir. H.M. (Michel) Bohms
Sr. Research Scientist
Structural Reliability

T +31 (0)88 866 31 07
M +31 (0)63 038 12 20
[hidden email]

Location

 

This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.

 

From: protege-user [mailto:[hidden email]] On Behalf Of Lieke Verhelst
Sent: donderdag 1 mei 2014 11:10
To: User support for WebProtege and Protege Desktop
Subject: Re: [protege-user] FW: advice on adding info to other people's ontology

 

Hi all,
Answering my own question to Joshua: provenance to be checked at e.g. the graph level.
Thanks for all your feedback..interesting discussion..
best, Lieke

> From: [hidden email]
> To:
[hidden email]
> Date: Thu, 1 May 2014 06:35:22 +0000
> Subject: Re: [protege-user] FW: advice on adding info to other people's ontology
>
> Hi Eric,
>
> Can you claify you point? (The whole issue is about difernt rdf:type's by different people)
>
> Thx Michel
>
> -----Original Message-----
> From: protege-user [
[hidden email]] On Behalf Of Eric Prud'hommeaux
> Sent: donderdag 1 mei 2014 6:54
> To: User support for WebProtege and Protege Desktop
> Subject: Re: [protege-user] FW: advice on adding info to other people's ontology
>
> dcterms:hasPart is typed as something other than owl:ObjectProperty
>
> * Bohms, H.M. (Michel) <
[hidden email]> [2014-04-30 14:29+0000]
> > Can we summarise (?):
> >
> > 1.
Best way: make 100% sure no issues: subclass own from source or reinvent (without subclassing)
> >
> > 2. Reasonable way: assume something that is somehow missing in the authorative source which is not contradicted by the source
> >
> > a. What would be IN THIS CASE the "unexpected result" by a reasoner?
> >
> > 3. Do not: assume things that likely contradict with explicit/implicit statements by the source (hats example)
>
> I think the only "safe" rule above is 1. That is, if someone accepts your verson of an ontology as a replacement for the authoritative copy, any changes to the original ontology can result in contradictions or overconstraints.
>
> 2 differs from 3 only in degree. If someone out there has data using dcterms:hasPart or dcterms:author as a DatatypeProperty, the modified ontology no longer fits their data, just as if the modified ontology were to say that all people where hats.
>
> It's really practical to create axioms which further constrain one's own data and people will be tempted to do this by patching ontologies. They have to evaluate whether it can cause unexpected consequneces with unsuspecting third parties as well as whether having a skewed reality will affect their ability to consume outside data.
>
>
> > This whole discussion does not only aply to dcterms:hasPart but also to the qudt units we reuse.
> > I assume they also fall in category 2.
Since the sources describes
> > several use cases where for each use case you might have to add some
> > stuff yourself first to make it work
> > (here: add the triple: <
http://qudt.org/schema/qudt#Unit>
> > rdfs:subClassOf rdfs:Datatype .)
> >
> >
> > Dr. ir. H.M. (Michel) Bohms
> > Sr. Research Scientist
> > Structural Reliability
> >
> > T +31 (0)88 866 31 07
> > M +31 (0)63 038 12 20
> > E
[hidden email]>
> >
> > Location<
http://www.tno.nl/locaties/DTM>
> >
> >
> >
> > [
<a href="cid:image001.gif@01CF6491.1F902690">cid:image001.gif@01CF6491.1F902690]<http://www.tno.nl/>
> >
> > This message may contain information that is not intended for you.
If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.
> >
> >
> >
> > From: protege-user [
[hidden email]] On
> > Behalf Of Ron Rudnicki
> > Sent: woensdag 30 april 2014 16:05
> > To: User support for WebProtege and Protege Desktop
> > Subject: Re: [protege-user] FW: advice on adding info to other
> > people's ontology
> >
> > If my caution applies at all, it applies only if DC publishes an .owl file in which dcterms:hasPart is typed as something other than owl:ObjectProperty.
If it does, then by changing the type in your file there could be queries that unexpectedly return results from your ontology but not from another that uses the canonical DC file or vice versa. In this case if you subtype, then you are warning your users that some extra effort will be needed to merge data, thereby saving us some debugging efforts. Of course, If DC offers no guidance on the implementation of hasPart, then typing it for one's own namespace is both legitimate and necessary.
> >
> > From: protege-user [
[hidden email]] On
> > Behalf Of Bohms, H.M. (Michel)
> > Sent: Wednesday, April 30, 2014 9:32 AM
> > To: User support for WebProtege and Protege Desktop
> > Subject: Re: [protege-user] FW: advice on adding info to other
> > people's ontology
> >
> > I assumed it works in MY/THIS case because I added a triple that was
> > actually in consideration of / in the spirit of the semantics intended
> > by DCI...(see argument Joshua) (so not being diff. semantics like in
> > your example with and without hats)
> >
> > So what you say is:
> > Make a new own hasPart being a subclass of dcterms:haPart and then add
> > the fact that my hasPart is of type owl:objectProperty To be even more cleaner and independent of my assumption that I actually model what was probably meant by DCI...
> >
> > ???
> >
> >
> >
> > Dr. ir. H.M. (Michel) Bohms
> > Sr. Research Scientist
> > Structural Reliability
> >
> > T +31 (0)88 866 31 07
> > M +31 (0)63 038 12 20
> > E
[hidden email]>
> >
> > Location<
http://www.tno.nl/locaties/DTM>
> >
> >
> >
> > [
<a href="cid:image001.gif@01CF6491.1F902690">cid:image001.gif@01CF6491.1F902690]<http://www.tno.nl/>
> >
> > This message may contain information that is not intended for you.
If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.
> >
> >
> >
> > From: protege-user [
[hidden email]] On
> > Behalf Of Ron Rudnicki
> > Sent: woensdag 30 april 2014 15:23
> > To: User support for WebProtege and Protege Desktop
> > Subject: Re: [protege-user] FW: advice on adding info to other
> > people's ontology
> >
> > But consider.
I take your class: Person and to it I add an axiom that Person is a subclass of things that wear hats. Now I publish my ontology claiming that the class person is the same as yours. I also collect instance data about Persons using my class. You try to aggregrate my data with yours, funny thing is that all these people wear hats...
> >
> > What I should have done instead was take your class as a parent class and create a subclass called PersonWithHat or HattedPerson. Then there's no problem.
> >
> > From: protege-user [
[hidden email]] On
> > Behalf Of Lieke Verhelst
> > Sent: Wednesday, April 30, 2014 9:03 AM
> > To:
> >
[hidden email]
> > >
> > Subject: [protege-user] FW: advice on adding info to other people's
> > ontology
> >
> >
> >
> > Hi all,
> >
> >
> >
> > Adding statements to an entity from a namespace that is not yours the
> > way
> >
> > Michel suggests is saying things about someone else's things like you
> >
> > were that other party.
It is assuming that this is technically possible.
> >
> > ( It isn't, because you cannot publish the statement in someone else's
> >
> > namespace since it is on someone else's server.)
> >
> > The editors allow it but imho this type of practice must be avoided
> >
> > because it causes unexpected results, just like Ron says. I haven't
> >
> > checked but hopefully reasoners act on the real origin of statements
> > on
> >
> > the web and ignore the statements about remote namespaces that are
> > locally
> >
> > declared.
> >
> >
> >
> > Of course it is perfectly fine to add statements in your own namespace
> >
> > that say something about someone else's entities.
> >
> >
> >
> >
> >
> > Best, Lieke
> >
> > Linked Data Factory
> >
> > Semantic web and Linked Data Services
> >
> > The Netherlands
> >
> >
> >
> > -----Original Message-----
> >
> > From: Bohms, H.M. (Michel)
> > [
[hidden email]]
> >
> > Sent: 29 April 2014 19:11
> >
> > To: '[hidden email]<
[hidden email]>'; 'User
> > support for WebProtege and Protege
> >
> > Desktop'
> >
> > Cc: 'Lieke Verhelst'
> >
> > Subject: RE: [protege-user] advice on adding info to other people's
> >
> > ontology
> >
> >
> >
> > Great, so I actually use it in the spirit of.
> >
> > I'll keep it the way I have it (with the triple added on my side)
> >
> >
> >
> > Thx Joshua for your clear analysis, Michel
> >
> >
> >
> > -----Original Message-----
> >
> > From: protege-user
> > [
[hidden email]] On
> >
> > Behalf Of Joshua TAYLOR
> >
> > Sent: dinsdag 29 april 2014 17:19
> >
> > To: User support for WebProtege and Protege Desktop
> >
> > Subject: Re: [protege-user] advice on adding info to other people's
> >
> > ontology
> >
> >
> >
> > On Tue, Apr 29, 2014 at 11:15 AM, Joshua TAYLOR
> > <
[hidden email]>>
> >
> > wrote:
> > If you have a property op_hasPart that's a subproperty of some DC property, then the DC property will still need to be declared as either an annotation property, a datatype property, or an object property, and the only one that makes sense here would be as an object property.
So this doesn't really avoid the problem.
> >
> > it is worth noting though, that the documentation about hasPart [1]
> >
> > says: "This term is intended to be used with non-literal values as
> > defined
> >
> > in the DCMI Abstract Model
> >
> > (
http://dublincore.org/documents/abstract-model/). As of December
> > 2007,
> >
> > the DCMI Usage Board is seeking a way to express this intention with a
> >
> > formal range declaration." It doesn't have a formal range, but the
> > fact
> >
> > that it's intended to be used with non-literal values at least means
> > that
> >
> > a declaration as an owl:ObjectProperty won't be too far from its
> > intended
> >
> > use. If nothing else, the values it will have as an object property
> > will
> >
> > be non-literals, so any triples with the owl:ObjectProperty hasParts
> > will
> >
> > be compatible the rdf:Property hasPart.
> >
> >
> >
> > //JT
> >
> >
> >
> > [1] http://dublincore.org/documents/dcmi-terms/#terms-hasPart
> >
> >
> >
> > --
> >
> > Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
> >
> > _______________________________________________
> >
> > protege-user mailing list
> >
> > [hidden email]
> > >
> >
> > https://mailman.stanford.edu/mailman/listinfo/protege-user
> >
> >
> >
> >
> >
> >
> >
> > Dit bericht kan informatie bevatten die niet voor u is bestemd. Indien
> > u
> >
> > niet de geadresseerde bent of dit bericht abusievelijk aan u is
> >
> > toegezonden, wordt u verzocht dat aan de afzender te melden en het
> > bericht
> >
> > te verwijderen. TNO aanvaardt geen aansprakelijkheid voor de inhoud
> > van
> >
> > deze e-mail, de wijze waarop u deze gebruikt en voor schade, van welke
> >
> > aard ook, die verband houdt met risico's verbonden aan het
> > elektronisch
> >
> > verzenden van berichten.
> >
> >
> >
> >
> >
> >
> >
> > This message may contain information that is not intended for you. If
> > you
> >
> > are not the addressee or if this message was sent to you by mistake,
> > you
> >
> > are requested to inform the sender and delete the message. TNO accepts
> > no
> >
> > liability for the content of this e-mail, for the manner in which you
> > use
> >
> > it and for damage of any kind resulting from the risks inherent to the
> >
> > electronic transmission of messages.
> >
> >
>
>
>
> > _______________________________________________
> > protege-user mailing list
> >
[hidden email]
> >
https://mailman.stanford.edu/mailman/listinfo/protege-user
>
>
> --
> -ericP
>
> office: +1.617.599.3509
> mobile: +33.6.80.80.35.59
>
> (
[hidden email])
> Feel free to forward this message to any list for any purpose other than email address distribution.
>
> There are subtle nuances encoded in font variation and clever layout which can only be seen by printing this message on high-clay paper.
> _______________________________________________
> protege-user mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-user
> _______________________________________________
> protege-user mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-user


_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user
Reply | Threaded
Open this post in threaded view
|

Re: FW: advice on adding info to other people's ontology

Eric Prud'hommeaux
In reply to this post by Bohms, H.M. (Michel)
* Bohms, H.M. (Michel) <[hidden email]> [2014-05-01 06:35+0000]
> Hi Eric,
>
> Can you claify you point? (The whole issue is about difernt rdf:type's by different people)

I can try, but I'm not sure which part was unclear. I'll annotate my points below.


> Thx Michel
>
> -----Original Message-----
> From: protege-user [mailto:[hidden email]] On Behalf Of Eric Prud'hommeaux
> Sent: donderdag 1 mei 2014 6:54
> To: User support for WebProtege and Protege Desktop
> Subject: Re: [protege-user] FW: advice on adding info to other people's ontology
>
> dcterms:hasPart is typed as something other than owl:ObjectProperty
>
> * Bohms, H.M. (Michel) <[hidden email]> [2014-04-30 14:29+0000]
> > Can we summarise (?):
> >
> > 1.      Best way: make 100% sure no issues: subclass own from source or reinvent (without subclassing)
> >
> > 2.      Reasonable way: assume something that is somehow missing in the authorative source which is not contradicted by the source
> >
> > a.      What would be IN THIS CASE the "unexpected result" by a reasoner?
> >
> > 3.      Do not: assume things that likely contradict with explicit/implicit statements by the source (hats example)
>
> I think the only "safe" rule above is 1. That is, if someone accepts your verson of an ontology as a replacement for the authoritative copy, any changes to the original ontology can result in contradictions or overconstraints.

1 is safe. Anything else is unsafe; example below. The text about "if someone accepts your version of an ontology" is intended to clarify the scope of the risk in changing someone else's ontology; it only affects those who use your version. For instance, if you say that foaf:Person rdfs:subClassOf my:PersonWithHat, and I never see that ontology or any data that was over-constrained by it, it won't affect me.


> 2 differs from 3 only in degree. If someone out there has data using dcterms:hasPart or dcterms:author as a DatatypeProperty, the modified ontology no longer fits their data, just as if the modified ontology were to say that all people where hats.

If you state foaf:Person rdfs:subClassOf my:PersonWithHat, you effectively over-constrain foaf:Person by partitioning the usual interpretation by whether the individual wheres a hat. Likewise, if you assert dcterms:hasPart rdf:type owl:ObjectProperty , you also partition the stated uses of dcterms:hasPart, rendering any statement like "fma:Body dctersm:hasPart 'kidney'." either inconsistent or some unspecified pun. What the designers intended is not so important as how publishers of related data and ontologies had interpreted it. In your case, since they pretty strongly indicated, albeit in ancillary documentation, that they intended the object to be a resource, there are probably rare and unimportant exceptions to your constraint (e.g. fma:Body dctersm:hasPart 'kidney').


> It's really practical to create axioms which further constrain one's own data and people will be tempted to do this by patching ontologies. They have to evaluate whether it can cause unexpected consequneces with unsuspecting third parties as well as whether having a skewed reality will affect their ability to consume outside data.
>
>
> > This whole discussion does not only aply to dcterms:hasPart but also to the qudt units we reuse.
> > I assume they also fall in category 2. Since the sources describes
> > several use cases where for each use case you might have to add some
> > stuff yourself first to make it work
> > (here: add the triple: <http://qudt.org/schema/qudt#Unit>
> > rdfs:subClassOf rdfs:Datatype .)
> >
> >
> > Dr. ir. H.M. (Michel) Bohms
> > Sr. Research Scientist
> > Structural Reliability
> >
> > T +31 (0)88 866 31 07
> > M +31 (0)63 038 12 20
> > E [hidden email]<mailto:[hidden email]>
> >
> > Location<http://www.tno.nl/locaties/DTM>
> >
> >
> >
> > [cid:image001.gif@01CF6491.1F902690]<http://www.tno.nl/>
> >
> > This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.
> >
> >
> >
> > From: protege-user [mailto:[hidden email]] On
> > Behalf Of Ron Rudnicki
> > Sent: woensdag 30 april 2014 16:05
> > To: User support for WebProtege and Protege Desktop
> > Subject: Re: [protege-user] FW: advice on adding info to other
> > people's ontology
> >
> > If my caution applies at all, it applies only if DC publishes an .owl file in which dcterms:hasPart is typed as something other than owl:ObjectProperty. If it does, then by changing  the type in your file there could be queries that unexpectedly return results from your ontology but not from another that uses the canonical DC file or vice versa. In this case if you subtype, then you are warning your users that some extra effort will be needed to merge data, thereby saving us some debugging efforts. Of course, If DC offers no guidance on the implementation of hasPart, then typing it for one's own namespace is both legitimate and necessary.
> >
> > From: protege-user [mailto:[hidden email]] On
> > Behalf Of Bohms, H.M. (Michel)
> > Sent: Wednesday, April 30, 2014 9:32 AM
> > To: User support for WebProtege and Protege Desktop
> > Subject: Re: [protege-user] FW: advice on adding info to other
> > people's ontology
> >
> > I assumed it works in MY/THIS case because I added a triple that was
> > actually in consideration of / in the spirit of the semantics intended
> > by DCI...(see argument Joshua) (so not being diff. semantics like in
> > your example with and without hats)
> >
> > So what you say is:
> > Make a new own hasPart being a subclass of dcterms:haPart and then add
> > the fact that my hasPart is of type owl:objectProperty To be even more cleaner and independent of my assumption that I actually model what was probably meant by DCI...
> >
> > ???
> >
> >
> >
> > Dr. ir. H.M. (Michel) Bohms
> > Sr. Research Scientist
> > Structural Reliability
> >
> > T +31 (0)88 866 31 07
> > M +31 (0)63 038 12 20
> > E [hidden email]<mailto:[hidden email]>
> >
> > Location<http://www.tno.nl/locaties/DTM>
> >
> >
> >
> > [cid:image001.gif@01CF6491.1F902690]<http://www.tno.nl/>
> >
> > This message may contain information that is not intended for you. If you are not the addressee or if this message was sent to you by mistake, you are requested to inform the sender and delete the message. TNO accepts no liability for the content of this e-mail, for the manner in which you use it and for damage of any kind resulting from the risks inherent to the electronic transmission of messages.
> >
> >
> >
> > From: protege-user [mailto:[hidden email]] On
> > Behalf Of Ron Rudnicki
> > Sent: woensdag 30 april 2014 15:23
> > To: User support for WebProtege and Protege Desktop
> > Subject: Re: [protege-user] FW: advice on adding info to other
> > people's ontology
> >
> > But consider. I take your class: Person and to it I add an axiom that Person is a subclass of things that wear hats. Now I publish my ontology claiming that the class person is the same as yours. I also collect instance data about Persons using my class. You try to aggregrate my data with yours, funny thing is that all these people wear hats...
> >
> > What I should have done instead was take your class as a parent class and create a subclass called PersonWithHat or HattedPerson. Then there's no problem.
> >
> > From: protege-user [mailto:[hidden email]] On
> > Behalf Of Lieke Verhelst
> > Sent: Wednesday, April 30, 2014 9:03 AM
> > To:
> > [hidden email]<mailto:[hidden email]
> > >
> > Subject: [protege-user] FW: advice on adding info to other people's
> > ontology
> >
> >
> >
> > Hi all,
> >
> >
> >
> > Adding statements to an entity from a namespace that is not yours the
> > way
> >
> > Michel suggests is saying things about someone else's things like you
> >
> > were that other party.  It is assuming that this is technically possible.
> >
> > ( It isn't, because you cannot publish the statement in someone else's
> >
> > namespace since it is on someone else's server.)
> >
> > The editors allow it but imho this type of practice must be avoided
> >
> > because it causes unexpected results, just like Ron says. I haven't
> >
> > checked but hopefully reasoners act on the real origin of statements
> > on
> >
> > the web and ignore the statements about remote namespaces that are
> > locally
> >
> > declared.
> >
> >
> >
> > Of course it is perfectly fine to add statements in your own namespace
> >
> > that say something about someone else's entities.
> >
> >
> >
> >
> >
> > Best, Lieke
> >
> > Linked Data Factory
> >
> > Semantic web and Linked Data Services
> >
> > The Netherlands
> >
> >
> >
> > -----Original Message-----
> >
> > From: Bohms, H.M. (Michel)
> > [mailto:[hidden email]<mailto:[hidden email]>]
> >
> > Sent: 29 April 2014 19:11
> >
> > To: '[hidden email]<mailto:'[hidden email]>'; 'User
> > support for WebProtege and Protege
> >
> > Desktop'
> >
> > Cc: 'Lieke Verhelst'
> >
> > Subject: RE: [protege-user] advice on adding info to other people's
> >
> > ontology
> >
> >
> >
> > Great, so I actually use it in the spirit of.
> >
> > I'll keep it the way I have it (with the triple added on my side)
> >
> >
> >
> > Thx Joshua for your clear analysis, Michel
> >
> >
> >
> > -----Original Message-----
> >
> > From: protege-user
> > [mailto:[hidden email]<mailto:protege-user-bo
> > [hidden email]>] On
> >
> > Behalf Of Joshua TAYLOR
> >
> > Sent: dinsdag 29 april 2014 17:19
> >
> > To: User support for WebProtege and Protege Desktop
> >
> > Subject: Re: [protege-user] advice on adding info to other people's
> >
> > ontology
> >
> >
> >
> > On Tue, Apr 29, 2014 at 11:15 AM, Joshua TAYLOR
> > <[hidden email]<mailto:[hidden email]>>
> >
> > wrote:
> > If you have a property op_hasPart that's a subproperty of some DC property, then the DC property will still need to be declared as either an annotation property, a datatype property, or an object property, and the only one that makes sense here would be as an object property. So this doesn't really avoid the problem.
> >
> > it is worth noting though, that the documentation about hasPart [1]
> >
> > says: "This term is intended to be used with non-literal values as
> > defined
> >
> > in the DCMI Abstract Model
> >
> > (http://dublincore.org/documents/abstract-model/). As of December
> > 2007,
> >
> > the DCMI Usage Board is seeking a way to express this intention with a
> >
> > formal range declaration."  It doesn't have a formal range, but the
> > fact
> >
> > that it's intended to be used with non-literal values at least means
> > that
> >
> > a declaration as an owl:ObjectProperty won't be too far from its
> > intended
> >
> > use.  If nothing else, the values it will have as an object property
> > will
> >
> > be non-literals, so any triples with the owl:ObjectProperty hasParts
> > will
> >
> > be compatible the rdf:Property hasPart.
> >
> >
> >
> > //JT
> >
> >
> >
> > [1] http://dublincore.org/documents/dcmi-terms/#terms-hasPart
> >
> >
> >
> > --
> >
> > Joshua Taylor, http://www.cs.rpi.edu/~tayloj/
> >
> > _______________________________________________
> >
> > protege-user mailing list
> >
> > [hidden email]<mailto:[hidden email]
> > >
> >
> > https://mailman.stanford.edu/mailman/listinfo/protege-user
> >
> >
> >
> >
> >
> >
> >
> > Dit bericht kan informatie bevatten die niet voor u is bestemd. Indien
> > u
> >
> > niet de geadresseerde bent of dit bericht abusievelijk aan u is
> >
> > toegezonden, wordt u verzocht dat aan de afzender te melden en het
> > bericht
> >
> > te verwijderen. TNO aanvaardt geen aansprakelijkheid voor de inhoud
> > van
> >
> > deze e-mail, de wijze waarop u deze gebruikt en voor schade, van welke
> >
> > aard ook, die verband houdt met risico's verbonden aan het
> > elektronisch
> >
> > verzenden van berichten.
> >
> >
> >
> >
> >
> >
> >
> > This message may contain information that is not intended for you. If
> > you
> >
> > are not the addressee or if this message was sent to you by mistake,
> > you
> >
> > are requested to inform the sender and delete the message. TNO accepts
> > no
> >
> > liability for the content of this e-mail, for the manner in which you
> > use
> >
> > it and for damage of any kind resulting from the risks inherent to the
> >
> > electronic transmission of messages.
> >
> >
>
>
>
> > _______________________________________________
> > protege-user mailing list
> > [hidden email]
> > https://mailman.stanford.edu/mailman/listinfo/protege-user
>
>
> --
> -ericP
>
> office: +1.617.599.3509
> mobile: +33.6.80.80.35.59
>
> ([hidden email])
> Feel free to forward this message to any list for any purpose other than email address distribution.
>
> There are subtle nuances encoded in font variation and clever layout which can only be seen by printing this message on high-clay paper.
> _______________________________________________
> protege-user mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-user
> _______________________________________________
> protege-user mailing list
> [hidden email]
> https://mailman.stanford.edu/mailman/listinfo/protege-user

--
-ericP

office: +1.617.599.3509
mobile: +33.6.80.80.35.59

([hidden email])
Feel free to forward this message to any list for any purpose other than
email address distribution.

There are subtle nuances encoded in font variation and clever layout
which can only be seen by printing this message on high-clay paper.
_______________________________________________
protege-user mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/protege-user