THIS SLIDE INTENTIONALLY LEFT BLANK

THIS SLIDE INTENTIONALLY LEFT BLANK

I H M C

DAML+OIL has no natural role for variables, and no variable binders. This may well be seen as an advantage:

(subclass A B) <==> (forall (?x)(implies (A ?x)(B ?x)))

P is restriction of Q on R to S <==> (forall (?x)(iff (P ?x)(and (Q ?x)(exists (?y)(and (R ?x ?y) (S ?y))))))

DAML+OIL Schema Parameter Interface Format

I H M C

However, many uses of, and likely future extensions to, DAML+OIL require the use of variables to allow pattern-matching:

1. allowing external software to bind values to parameters (CoABS policy tool; DAML-S applications)

2. syntax for atomic component of 'rules' and queries

3. controlling use of URL's across networks (as in HORUS)

DAML+OIL Schema Parameter Interface Format

I H M C

Changing DAML to include variables raises a host of semantic issues which we propose to completely avoid. This is a purely syntactic proposal. No semantics is provided for the variables until they have been completely instantiated. In logical terms, these are schematic (not quantified) variables; in programming language terms, they are macros. To avoid confusion we call them parameters.

DAML+OIL distinguishes two kinds of 'name': URIs and literals. The proposal is an extended notation in which URI's may be replaced by URI parameters and literals by literal parameters.

DAML+OIL Schema Parameter Interface Format

I H M C

Uri-parameter ::= ?word /Literal-parameter ::= ??word /word ::= [^< > ? / space ]+

<Or some other syntax would do>

A URI [literal] term is either a URI [literal] or a URI [literal] parameter. A substitution is a sequence of substitution pairs consisting of a URI [literal] parameter and a URI [literal] term.

DAML+OIL Schema Parameter Interface Format

I H M C <daml:Class rdf:ID=“?actionType/” > <rdfs:label>?actionType/</rdfs:label> <rdfs:subClassOf rdf:resource="#Ordinary Action"/> <rdfs:subClassOf> <daml:Restriction daml:Cardinality="1”> <daml:onProperty rdf:resource="#performedBy"/> <daml:toClass rdf:resource="?performer/"/> </daml:Restriction> </rdfs:subClassOf></daml:Class>

DAML+OIL Schema Parameter Interface Format

I H M C <daml:Class rdf:ID=“?actionType/” > <rdfs:label>?actionType/</rdfs:label> <rdfs:subClassOf rdf:resource="#Ordinary Action"/> <rdfs:subClassOf> <daml:Restriction daml:Cardinality="1”> <daml:onProperty rdf:resource="#performedBy"/> <daml:toClass rdf:resource=”?performer/"/> </daml:Restriction> </rdfs:subClassOf></daml:Class>

?actionType/ #DelegateAction?performer/ #Domain_Manager

?actionType/ #Multi_Actor_Action?performer/ #ActorGroup

DAML+OIL Schema Parameter Interface Format

I H M C <rdfs:Class rdf:ID="elephant">

<rdfs:subClassOf>

<rdfs:Class rdf:about="#animal"/>

</rdfs:subClassOf>

<rdfs:subClassOf>

<daml:Restriction>

<daml:onProperty rdf:resource="#eats"/>

<daml:toClass>

<rdfs:Class rdf:about="#plant"/>

</daml:toClass>

</daml:Restriction>

</rdfs:subClassOf>

<rdfs:subClassOf>

<daml:Restriction>

<daml:onProperty rdf:resource="#colour"/>

<daml:hasValue>

<daml:ConcreteTypeExpression>EQUAL ``grey''

</daml:ConcreteTypeExpression>

</daml:hasValue>

</daml:Restriction>

</rdfs:subClassOf>

</rdfs:Class>

A more compact notation (with Frank v. Harmelen)

Class elephant

SubClassOf #animal

[which #eats #plant;

which #color EQUAL `grey’]

I H M C <rdfs:Class rdf:ID="herbivore"> <daml:intersectionOf> <rdfs:Class rdf:about="#animal"/> <daml:Restriction> <daml:onProperty rdf:resource="#eats"/> <daml:toClass>

<rdfs:Class> <daml:unionOf> <rdfs:Class

rdf:about="#plant"/> <daml:Restriction> <daml:onProperty

rdf:resource="#is_part_of"/> <daml:hasValue>

<rdfs:Class rdf:about="#plant"/>

</daml:hasValue> </daml:Restriction> </daml:unionOf></rdfs:Class>

</daml:toClass> </daml:Restriction> </daml:intersectionOf></rdfs:Class>

Class herbivore

IntersectionOf

[#animal which #eats

UnionOf

[#plant;

which #is_part_of hasValue #plant] ]

A more compact notation (with Frank v. Harmelen)

I H M C

A brief essay inspired by conversations with Lynn SteinBill AndersenJim Hendler Adam PeaseJerry Hobbs and others too numerous to mention

I H M C

<rdf:RDF xmlns:rdf =http://www.w3.org/1999/02/22-rdf-syntax-ns# xmlns:a =http://someplace.org/this_ontology# xmlns:b=http://some_other_place.gov/that_ontology#

<daml:Class rdf:ID=”Barble"><rdfs:subClassOf> <daml:Restriction> <daml:onProperty rdf:resource=”a#foodle"/> <daml:toClass rdf:resource=”b#bazset"/> </daml:Restriction> </rdfs:subClassOf></daml:Class>

I H M C

If I use a term from your ontology, how much of your ontology am I assuming?

I H M C

If I use a term from your ontology, how much of your ontology am I assuming? A1: Only the parts that I explicitly endorse or include

I H M C

If I use a term from your ontology, how much of your ontology am I assuming? A1: Only the parts that I explicitly endorse or includeA2: Enough to establish the intended meaning of the term

I H M C

If I use a term from your ontology, how much of your ontology am I assuming? A1: Only the parts that I explicitly endorse or includeA2: Enough to establish the intended meaning of the termA3: All of it

I H M C

If I use a term from your ontology, how much of your ontology am I assuming? A1: Only the parts that I explicitly endorse or includeA2: Enough to establish the intended meaning of the termA3: All of itA4: None of it

I H M C

If I use a term from your ontology, how much of your ontology am I assuming? Wrong answer: The question doesn’t arise since we are all basically part of One World Wide Ontology

On

tolo

gy

I H M C

OK, at least we are using the same notation, which is a good start.

I H M C

OK, at least we are using the same notation, which is a good start.

BUT if you are using a closed world assumption and I am not, we will misunderstand one another.

I H M C

OK, at least we are using the same notation, which is a good start.

BUT

I H M C

You say… Employees are a subclass of personsI say… Employment is a property some people have

I H M C

You say… Things are physical objectsI say… Things include ideas, integers, intentions, states of mind, The Grinch…

I H M C

You say… Things are physical objectsI say… Things include ideas, integers, intentions, states of mind, The Grinch, fluents, tropes, eventualities, ways of being, possible worlds, whatever

I H M C

You say… Things are physical objectsI say… Things are physical objects

I H M C

You say… Things are physical objectsI say… Things are physical objectsBUTYou say… Physical objects retain their identity through time and change their propertiesI say… Physical objects have temporal parts which have timeless properties

I H M C

You say… Things are physical objectsI say… Things are physical objectsBUTYou say… Physical objects retain their identity through time and change their propertiesI say… Physical objects have temporal parts which have timeless properties

Describing events/actions/processes/occurrents/happenings…

Things change…so need to make assertions relative to times

TrickleDown: push temporal parameters as deep into the expression as they need to go (but no further).

R(a,b,c) + T

[R(a,b,c),T] simple proposition true at time

[R,T](a,b,c) time-indexed relation (eg position)

R(a,b,c,[T]) time-relative assertion

R([a,T],b,[c,T]) dynamic entities

Often most natural language for humans

Often most effective for automatic reasoning about complex domain

Often most effective for action planning

Corresponds to linguistic analyses

I H M C

You say… holes are physical things, and holes can only be wholeI say… holes are spatio-temporal but not physical, and can have partsHe says… holes aren’t physical, but are necessarily located in physical things and cannot have partsAnd he says… holes don’t exist as entities, but are locations which are empty.

I H M C

Which ontology should be chosen/adopted/made into the standard?

I H M C

Which ontology should be chosen/adopted/made into the standard? WRONG QUESTION !All these viewpoints have their uses, are internally coherent and are legitimate points of view. There is no single best answer.

I H M C

Long TermWays to broker meanings between different ontological perspectives (translation techniques, known safe routes, neutral ontological territory.)Medium TermWays to SAY what parts of one ontology are being included into another, and to analyze the consequences.Short TermAgree that there is a problem to be solved. LUNCH