a model-driven semantic web david frankel (david frankel consulting) pat hayes ( institute for human...

18
A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes (Institute for Human & Machine Cognition, University of West Florida) Elisa Kendall (Sandpiper Software) Deborah McGuinness (Knowledge Systems Laboratory, Stanford University)

Upload: cleopatra-allen

Post on 28-Dec-2015

216 views

Category:

Documents


2 download

TRANSCRIPT

Page 1: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web

David Frankel (David Frankel Consulting)

Pat Hayes (Institute for Human & Machine Cognition, University of West Florida)

Elisa Kendall (Sandpiper Software)

Deborah McGuinness (Knowledge Systems Laboratory, Stanford University)

Page 2: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web 2

Metadata Management Scenario1—Integrated MOF Repository

Page 3: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web 3

Metadata Management Scenario 2—Federated MOF Repositories

Page 4: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web 4

MOF and the Semantic Web

• MOF technology streamlines the mechanics of managing models as XML documents, Java objects, CORBA objects, etc.

• Semantic Web technology is for reasoning about resources– Reasons over models that describes the resources

• MOF provides no help with reasoning• Semantic Web not focused on helping with

mechanics of managing models• Complementary technologies

– Despite some overlap

Page 5: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web 5

Applications

• Using ontology-based reasoning to:– Detect problematic financial transactions– Analyze complex business rules– Resolve multiple policy constraints in

• Runtime applications management

• Network management

• While integrating with: – Pre-existing relational and OLAP data stores– Entity-relationship models– UML models– etc.

Page 6: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web 6

Business Information Model

context PreferredChecking inv:

--Cannot go below the minBalance

balance >= minBalance

CheckingAccount<<BusinessEntity>>

minBalance : Money

Account<<BusinessEntity>>

id : Stringbalance : Money

Customer<<BusinessEntity>>

socialSecurityNum : Stringname : Stringaddress : String

1..n

1

+account

1..n

+customer1

PreferredChecking<<BusinessEntity>>

{disjoint}

Money<<DataType>>

SavingsAccount<<BusinessEntity>>

interestRate : Decimal

Decimal<<DataType>>

Page 7: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web 7

Business Service Model

FundsXFer<<BusinessService>>

XFerFromChecking(in fromAcct : CheckingAccount, in toAcct : SavingsAccount, in amount : Money)

context FundsXFer::XFerFromChecking (fromAcct : CheckingAccount, toAcct : SavingsAccount) : voidpre: --There must be sufficient funds in the checking account to support the transfer fromAcct.balance >= amountpre: --The checking account and the savings account must belong to the same customer fromAccount.customer = toAccount.customer

post: --The balance of the checking account is reduced from its orginal amount by the amount of the transfer fromAcct.balance = fromAcct.balance@pre - amountpost: --The balance of the savings account is increased from its original amount by the amount of the transfer toAcct.balance = toAcct.balance@pre + amount

Page 8: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web 8

UML and the Semantic Web

• No established graphical notation for Semantic Web languages

• UML profiling is one approach

• UML profiles for RDF, OWL, and Topic Maps

Page 9: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web 9

Toward a Model-Driven Semantic Web

• Submission team– IBM– Sandpiper Software– Gentleware– AT&T (Supporter)– DSTC

• MOF metamodels models of – RDF/S – OWL– Simplified Common Logic– Topic Maps– Entity-Relationship

• UML Profiles for RDF/S, OWL• Mappings

– UML-OWL – (UML-RDF/S)– (UML-Topic Map)

Page 10: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web 10

Bridging the Semantic Web and MDA Worlds

MOF/XMI Based Tooling (e.g. Semantic Web Eclipse Plug-In)

MOF/XMI Based Tooling (e.g. Semantic Web Eclipse Plug-In)

OWL-XMI Bridge

OWL-XMI Bridge

Native Semantic WebOntology Development

Tool

Native Semantic WebOntology Development

Tool

Generic UML Modeling Tool

Generic UML Modeling Tool

XMI OWL Document

XMI OWL DocumentNative

OWL Document

Native OWL Document

UML ModelUsing UML Profile

UML ModelUsing UML Profile

UML-OWL Bridge

UML-OWL Bridge

GenericUML Model

GenericUML Model

A B C

Means tool A outputs B, and B serves as input to tool C.

Page 11: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web 11

Translating Entity Relationship Models to OWL

MOF/XMI Based Tooling (e.g. Semantic Web Eclipse Plug-In)

MOF/XMI Based Tooling (e.g. Semantic Web Eclipse Plug-In)

OWL-XMI Bridge

OWL-XMI Bridge

Native Semantic WebOntology Development

Tool

Native Semantic WebOntology Development

Tool

Generic UML Modeling Tool

Generic UML Modeling Tool

XMI OWL Document

XMI OWL DocumentNative

OWL Document

Native OWL Document

UML ModelUsing UML Profile

UML ModelUsing UML Profile

UML-OWL Bridge

UML-OWL Bridge

GenericUML Model

GenericUML Model

A B C

Means tool A outputs B, and B serves as input to tool C.

ER ModelER Model

ER ToolER Tool

ER-OWL Bridge

ER-OWL Bridge

Page 12: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web 12

Metamodel Contents

• Model of abstract syntax– Defines a domain of conforming abstract

syntax trees

OWLMetamodel

OWLMetamodel

Abstract Syntax

OWLOntology

OWLOntology

Abstract Syntax Tree

= conforms to

SCLMetamodel

SCLMetamodel

Abstract Syntax

SCLExpression

SCLExpression

Abstract Syntax Tree

Page 13: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web 13

Part of the SCL Abstract Syntax

CommentedTerm

comment : String

SequenceVariable

Term

1

0..1

1

0..1

Application

0..1

0..1

0..1

0..1

0..*

0..1

+functionPositionTerm0..*{ordered}

0..1

1

0..1

+additonalTerm1

0..1

Name

string : String

'('quotedstring S term S? ')' '(' S? term S termseq S? ')'

termseq ::= (term S?)* seqvar?term ::= name | '(' S? term S termseq S? ')' | '('quotedstring S term S? ')'

Page 14: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web 14

Representation of an Abstract Syntax Tree

functionPositionTerm

(married Jack Jill)

: Application

Married : Name Jack : Name Jill : Name

additionalTerm additionalTerm

Page 15: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web 15

ODM MetamodelODM Metamodel

Concrete XML-Based Syntax for CWM(XML DTD or

Schema)

Concrete XML-Based Syntax for ODM(XML Schema)

Abstract Syntax

Mapping

XMI: The MOF-XML Technology Mapping

Page 16: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web 16

Corresponding Part of the XMI Schema

<xsd:complexType name="Application">

<xsd:complexContent>

<xsd:extension base="omgorg.odm.scl:Term">

<xsd:choice maxOccurs="unbounded" minOccurs="0">

<xsd:element name="sequenceVariable"

type="omgorg.odm.scl:SequenceVariable"/>

<xsd:element name="functionPositionTerm"

type="omgorg.odm.scl:Term"/>

<xsd:element name="additonalTerm"

type="omgorg.odm.scl:Term"/>

</xsd:choice>

</xsd:extension>

</xsd:complexContent>

Page 17: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web 17

The Core Metamodel

DL Core MetamodelDL Core Metamodel

UML MetamodelUML Metamodel

Means bi-directional mapping

OWL/RDF/S MetamodelOWL/RDF/S Metamodel

ER MetamodelER Metamodel Topic Maps MetamodelTopic Maps Metamodel

Page 18: A Model-Driven Semantic Web David Frankel (David Frankel Consulting) Pat Hayes ( Institute for Human & Machine Cognition, University of West Florida) Elisa

A Model-Driven Semantic Web 18

Synergy Summary

• MOF’s model management facilities and the Semantic Web’s reasoning capabilities are separate, complementary concerns

• The ability of Semantic Web reasoners to find discrepancies in invariant rules, preconditions, and postconditions, can add scalability to MDA’s use of Design-by-Contract (DBC)

• UML profiles can serve as graphical notations for Semantic Web languages