Hey Dave, (01)
this is exactly what i tried to express in my first mail. It's about
extensional and intensional (edit: definition of) sets, yes! :) And
OWL/DL employs both, thats what i needed to know. I had the feeling that
there was something addional that OOP cant do. This is the first part of
the puzzle. (02)
@Alex: yes :)) thx for the advice!
@Patrick: Ok i understand what you are saying. I suppose this is the
viewpoint of a deducer and of course therefor the semantics of an
ontology. Thx for that.
@Ingvar: Thx for the paper. I started to read it and its interesting,
but there are too many concepts that i'm not aware of. But interesting
site. Someday i will read some philosophy texts from there. Thx.
@Duane: Your personal opinion made me smile. :) Not that i can judge
this, but the imagination is funny.
@Rich: Reminded me that ontologies are logic, not data structures. :D
@Christopher: Ok, i get it, it's about the definitions of sets, not the
sets themselves. This is like what i read in the paper of Ingvar:
realist mode of speech vs conceptual mode of speech. :D (03)
So can i say, intersection is intensional definition and "ordinary"
subclassing is extensional definition? Hm no?! Christopher wrote, that
an extensional defined set just lists all its members and an intensional
defined set says, how its members should look like by constraining its
properties. Ok, but then i'm not sure how either of these look like. How
does each look like in OWL? (04)
Duane wrote, that multiple classes are involved in an intersection. If i
have a look at the wine ontology example
(http://www.w3.org/TR/2004/REC-owl-guide-20040210/wine.rdf) this seems
to fit. But there are also many intersections made of one class and one
or more restrictions: (05)
<owl:Class rdf:ID="WhiteWine">
<owl:intersectionOf rdf:parseType="Collection">
<owl:Class rdf:about="#Wine"/>
<owl:Restriction>
<owl:onProperty rdf:resource="#hasColor"/>
<owl:hasValue rdf:resource="#White"/>
</owl:Restriction>
</owl:intersectionOf>
</owl:Class> (06)
Whats up with that? Seems to me like subclassing. (07)
And even more confusing: (08)
<owl:Class rdf:ID="Sauternes">
<rdfs:subClassOf rdf:resource="#LateHarvest"/>
<rdfs:subClassOf rdf:resource="#Bordeaux"/>
..................
</owl:Class> (09)
Whats up with that? Seems to me like intersecting. :D (010)
So can i do what i want in an ontology? It doesnt matter if i use
subclass-tag or intersection-tag, the result is the same? Or where is
the semantic difference? What if i declared WhiteWine as subclass and
Sauternes as intersection? Thats why i came up with my explanations
involving properties, because i tried to explain those things in the
example ontology somehow. I think i'm still confused. :) (011)
Best regards to all and thx for the replies!
Mario Uhlig (012)
Dave McComb schrieb:
> Garz (and anyone else following this thread)
>
> We teach a class in OWL. My sense from our students confusion is that the
>main point of departure between what you understand and what's occurring in an
>ontology comes down to the distinction between intensional and extensional
>sets. An extensional set has it's members individually assigned to the set.
>This is how almost all of OO and Relational works. People are in the set of
>employees because you put them in the employee master table or your program
>did a new method on the employee class.
>
> An intensional set is one where you get members via some sort of inclusion
>criteria. OWL/DL employs both strategies. Your interface strategy is taking
>advantage of the intensional definition (although as in OO you could also
>asert something to be a member of the intersection)
>
> _____
> From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
>[ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Rich Cooper
>[rich@xxxxxxxxxxxxxxxxxxxxxx]
> Sent: Sunday, July 26, 2009 9:25 AM
> To: '[ontolog-forum] '
> Subject: Re: [ontolog-forum] When to use pure subclassing and when to use
> subclassing by intersecting
>
> It seems to me that the major difference is that logic presumes a global
>dictionary, so all properties are specifically defined and associated with a
>globally unique identifier. OOP classes have private symbols, and the meaning
>of symbols (read, write, …) depends on which class(es) interpret the symbols.
>In OOP, only the type names are globally uniquely defined. All others are
>locally defined within a scope.
>
> -Rich
>
>
>
> Sincerely,
> Rich Cooper
> EnglishLogicKernel.com
> Rich AT EnglishLogicKernel DOT com
>
> ________________________________
> From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx
>[mailto:ontolog-forum-bounces@xxxxxxxxxxxxxxxx] On Behalf Of Duane Nickull
> Sent: Sunday, July 26, 2009 8:18 AM
> To: [ontolog-forum]
> Subject: Re: [ontolog-forum] When to use pure subclassing and when to use
>subclassing by intersecting
>
> Agree here. As a programmer first, the subclass relationship has special
>meaning.
>
> //AS3 super class
> Private class SuperClass():void {
> public var name:String;
>
> //methods
> private setName (String name) {
> name = this.name;
> }
>
> private getName():void {
> returns name;
> }
> }
>
> And now a subclass that extends the super
>
> Private class SubClass():void extends SuperClass{
> // this should do everything the super does but add something new.
> // by using “extends” it will inherit the Superclasses methods.
>
> public var age:Integer;
>
> //methods. Using uint instead of int since age will normally be 0 >
> private setAge (uint age) {
> age = this.age;
> }
>
> private getAge():void {
> returns age;
> }
> }
>
> Now an intersection class is often used in UML where CEP principles are
>involved. This is however not a true intersection so it is usually called an
>Association class. Association class In modeling an association, there are
>times when you need to include another class because it includes valuable
>information about the relationship. For this you would use an association
>class that you tie to the primary association. An association class is
>represented like a normal class. The difference is that the association line
>between the primary classes intersects a dotted line connected to the
>association class. An example might be a class called “Event” and another one
>called “Entity” and the association class might document how the two are
>related (An Entity can be a catalyst for an event).
>
> A true intersection must take two classes unrelated by their hierarchies'
>(often difficult since most taxonomies of classes in programming all inherit
>from the same primitive. For example, everything in Java inherits from the
>Object:object). There are some times when these do exist though. Imagine you
>have a HusbandNetWealth and a WifeNetWealth class, you could say there is an
>intersection called “MarriedCoupleNetWealth” however many OO languages allow
>for something called Multiple inheritance and polymorphisms. The latter class
>in this case can inherit from both the previous classes therefore rendering
>the intersection class notion completely worthless for most programmers and
>they simply regard it as an unnecessary construct invented by people who do
>not understand class inheritance.
>
> Strictly speaking, from a programming perspective, most people will try to
>maintain some lineage of a hierarchy rather than trying to make an
>“intersection class” which is why ontologists who do this have a hard time
>trying to find programmers willing to implement their work.
>
> Of course, this is all my personal opinion and others on the list will
>probably argue against it.
>
> Duane Nickull
>
> On 7/25/09 7:49 AM, "garz" <garz@xxxxxxx> wrote:
> Hi Patrick,
>
> i'm an object-oriented programmer, maybe thats why i have problems
> understanding the concepts of an ontology. I dont want to model
> anything, i want to write some components that enable you to create and
> use ontologies in a some program. Therefor i need to understand the
> concepts, so that those components can offer the full spectrum and in a
> convenient way.
>
> In object-oriented programming there are also classes and inheritance.
> There i subclass a class so that i can add new properties to it (besides
> adding new methods, which arent relevant for ontologies). You write,
> that i cant have properties in a subclass that arent also in the
> parentclass, but instances of a subclass can have properties that
> instances of the superclass dont have because those properties arent
> required in the superclass, but are required in the subclass. Is that right?
>
> my question arose by studying the two example ontologies food.rdf and
> wine.rdf in the ontology guide on the w3-site:
> http://www.w3.org/TR/2004/REC-owl-guide-20040210/food.rdf
> http://www.w3.org/TR/2004/REC-owl-guide-20040210/wine.rdf
>
> in the food.rdf, there are the classes:
> <owl:Class rdf:ID="ConsumableThing"/>
>
> <owl:Class rdf:ID="PotableLiquid">
> <rdfs:subClassOf rdf:resource="#ConsumableThing"/>
> <owl:disjointWith rdf:resource="#EdibleThing"/>
> </owl:Class>
>
> and in the wine.rdf there is the class:
> <owl:Class rdf:ID="Wine">
> <rdfs:subClassOf
>
>rdf:resource="http://www.w3.org/TR/2003/PR-owl-guide-20031209/food#PotableLiquid"/<http://www.w3.org/TR/2003/PR-owl-guide-20031209/food#PotableLiquid>>
> <rdfs:subClassOf>
> <owl:Restriction>
> <owl:onProperty rdf:resource="#hasMaker"/>
> <owl:cardinality
>
>rdf:datatype="http://www.w3.org/2001/XMLSchema#nonNegativeInteger";>1</owl:cardinality<http://www.w3.org/2001/XMLSchema#nonNegativeInteger>>
> </owl:Restriction>
> </rdfs:subClassOf>
> .....
> </owl:Class>
>
> <owl:ObjectProperty rdf:ID="hasMaker">
> <rdf:type
>rdf:resource="http://www.w3.org/2002/07/owl#FunctionalProperty"/<http://www.w3.org/2002/07/owl#FunctionalProperty>>
> </owl:ObjectProperty>
>
> Here, the class Wine defines new properties, that the class
> PortableLiquid doesnt posses. But this is done via anonymous classes
> that are defined between the <rdfs:subClassOf>-tags? So the Wine class
> is subclass of PortableLiquid and a anonymous class that defines the
> property hasMaker and that it this property required, right? So to say,
> the new properties that the Wine class posseses are tricked into the
> Wine class? Tricked because subclasses cant have new properties, like
> you wrote respectively like i hope to have understood it in the right way.
>
> I want to very much thank you, both Patrick and Alex, for taking your
> time to answer my questions.
> @Alex, i will come back to your explanation, as soon as i understand a
> little more.
>
> Best regards
> Mario Uhlig
>
>
> Patrick Cassidy schrieb:
>
>> Garz,
>> I am not quite clear about the question, and the distinctions you are
>> making. Every subclass merely picks out of some parent class instances
>> that have specific properties. So you cannot have in a subclass properties
>> that are not also in the parent class, but you can *require* that the
>> instances of the subclasses have a specific property, even if that property
>> is not "mentioned" (i.e. not required) for every instance of the parent
>> class.
>>
>> The term "intersection" is usually used to refer to an intersection of
>> classes - the intersection is the class of entities, each of which is an
>> instance of both parent classes. That intersection is a subclass of each of
>> the parent classes. If you have a class of "RedThings" (anything that is
>> red) and a class of "Wines" and you define a class of "RedWines" it could be
>> a subclass of both RedThings and of "Wines", and would be the same as the
>> intersection of "RedThings" and "Wines", unless you also add other
>> restrictions to the "RedWines" class. [[ this is a bad example, because
>> colors are too heterogeneous in application for "RedThings" to be a useful
>> class]]. You seem to be using "intersection" in a different sense.
>>
>> Perhaps you could give a more specific example of what you would like to
>> do.
>>
>> Pat
>>
>> Patrick Cassidy
>> MICRA, Inc.
>> 908-561-3416
>> cell: 908-565-4053
>> cassidy@xxxxxxxxx
>>
>>
>>
>>
>>> -----Original Message-----
>>> From: ontolog-forum-bounces@xxxxxxxxxxxxxxxx [mailto:ontolog-forum-
>>> bounces@xxxxxxxxxxxxxxxx] On Behalf Of garz
>>> Sent: Friday, July 24, 2009 12:35 PM
>>> To: ontolog-forum@xxxxxxxxxxxxxxxx
>>> Subject: [ontolog-forum] When to use pure subclassing and when to use
>>> subclassing by intersecting
>>>
>>> Hi,
>>>
>>> i'm trying to create an ontology and there is a problem that i have in
>>> understanding the difference between a pure subclass and a subclass
>>> defined as intersection. my own explanation for this is the following:
>>>
>>> a subclass defined as intersection relates to the properties the
>>> superclass has and constraints them. for example redwine constraints
>>> the
>>> property color of the superclass wine by constraining its value to red.
>>> redwine is like an aspect or a view on a class, in this case on the
>>> aspect color. or its like an alias picking a subset out of the related
>>> superclass. a class defined that way introduces a new concept or term
>>> by
>>> just using the already existing, its not capable of creating new
>>> content, so to speak.
>>>
>>> a pure subclass should then be used, if there is something new to it,
>>> represented as a property that isnt mentioned in the superclass and the
>>> subclass is free to introduce this property or not. if the property
>>> already was in the superclass, a subclass defined by intersection would
>>> be sufficient.
>>>
>>> am i correct with this thinking? i dont really get it. or is
>>> intersection always applicable, but in this case i really dont
>>> understand the sense behind it. plz enlighten me.
>>>
>>> regards
>>>
>>> _________________________________________________________________
>>> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
>>> Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
>>> Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
>>> Shared Files: http://ontolog.cim3.net/file/
>>> Community Wiki: http://ontolog.cim3.net/wiki/
>>> To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
>>> To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx
>>>
>>>
>>>
>> _________________________________________________________________
>> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
>> Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
>> Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
>> Shared Files: http://ontolog.cim3.net/file/
>> Community Wiki: http://ontolog.cim3.net/wiki/
>> To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
>> To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx
>>
>>
>>
>>
>
>
> _________________________________________________________________
> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
> Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
> Shared Files: http://ontolog.cim3.net/file/
> Community Wiki: http://ontolog.cim3.net/wiki/
> To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
> To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx
>
>
> _________________________________________________________________
> Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
> Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
> Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
> Shared Files: http://ontolog.cim3.net/file/
> Community Wiki: http://ontolog.cim3.net/wiki/
> To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
> To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx
>
>
> (013)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (014)
|