Enhancing the Semantics Management Capabilities of Metadata Registries

1
Enhancing the Semantics Management Capabilities
of Metadata Registries Wuhan Symposium on
Ontology and MetaModeling Wuhan University, P. R.
China March 16-17, 2006
Kevin D. Keck Lawrence Berkley National
Laboratory http//sdm.lbl.gov/kdkeck/ kdkeck_at_lbl.
gov
2
Just In Time Slides

• These slides are substantially modified from the
  version printed in the program
• (just following Bruce's example)
• tinyurl for XMDR presentations page
• http//tinyurl.com/z7aun

3
Topics

• Integrating Ontology Management, Vocabulary
  Management, and Data Management
• Implementation for the Semantic Web
• Deployment and Best Practices

4

• Integrating
• Ontology Management, Vocabulary Management,
• and Data Management

5
Historically Distinct Disciplines

• Data Management
• ISO SQL
• OMG UML, CWM
• W3C XML
• Vocabulary Management
• ISO TC37
• OMG TQS
• W3C SKOS
• Ontology Management
• ISO Common Logic
• OMG ODM
• W3C OWL/RDF

6
Aligning Standards 11179 19763 ( MOF ODM)
ISO/IEC19763
ISO/IEC 11179-2
XMDR Registry
Terminology Basic Classes Basic
Relationship
Ontologies
Analysis and Extraction
Registering
7
Infrastructure is Expensive

• Developing, deploying, and maintaining distinct
  software for each kind of repository is
  time-consuming and expensive
• both on the server side, and on the client side
• Common architectural challenges motivate reuse of
  technologies
• scalability, reliability, auditability, etc.

8
Over Time, Content IsEven More Expensive

• Common registries facilitate reuse of content
• Coordination (maintenance) is easier within an
  integrated registry system
• Mature registry services can maximize content
  availability (through multiple APIs)
• Mature infrastructure will protect content
  investments

9
Bridging different realms of metadata standards
11179 Edition 3 Metadata Registry 19763
Framework for Metamodel Interoperability
10
11179 Designation and Definition
11
Data Models vs.Conceptual Models
MOF Core Ordered, Hierarchical Containment
Concept System Flat Semantic Network
12
Twist between Data Elements and Concepts

• Data Elements do not correspond to a single
  Concept, but rather a pair of Concepts.
• This is an intrinsic characteristic of
  informationit is always an assertion of some
  predicate, for some subject.

13
Concepts and Relationships
14
Concept System Draft One
15
Concept System Draft Two
16
Concept System
17
Ontology
18

• Implementation
• for the
• Semantic Web

19
XMDR files serve dual purposeXML and
OWL-compatible RDF

• Conform to XMDR XML schema
• More human-readable
• Easier to manipulate with XML tools, such as XSLT
• XML serialization of RDF
• Base tag includes rdfabout attribute
• Literals encoded as element content
• URIs encoded as attribute values
• striped resource, property, resource, use
  abbreviated form for anonymous nodes
• Conform with 11179 OWL ontology

20
XMDR Prototype Example dual purposeRDF/XML
file DEALL.1.5394.1.xml

• ltDataElement rdfabout""
• xmlbase"http//xmdr.lbl.gov/xmdr/data/DEALL.1.53
  94.1.xml"gt
• ltcontainer rdfresource"http//oaspub.epa.gov/e
  dr"/gt
• ltidentifier rdfparseType"Resource"gt
• ltstring rdfdatatype"xsdstring"gt5394lt/strin
  ggt
• lt/identifiergt
• ltversion rdfdatatype"xsdstring"gt1lt/versiongt
• ltadministrationRecord rdfparseType"Resource"gt
• ltregistrationStatus rdfdatatype"xsdstring"
  gtStandardlt/registrationStatusgt
• ltadministrativeStatus rdfdatatype"xsdstrin
  g"gtFinallt/administrativeStatusgt
• ltcreationDate rdfdatatype"xsddate"gt1999-09
  -09lt/creationDategt
• lt/administrationRecordgt
• ltdesignation rdfparseType"Resource"gt
• ltcontext rdfresource"CXT-Legacy.xml"/gt
• ltsign xmllang"en"gtCountry Namelt/signgt
• lt/designationgt
• ltdesignation rdfparseType"Resource"gt
• ltcontext rdfresource"CXT-Long
  Abbreviation.xml"/gt
• ltcontext rdfresource"CXT-Medium
  Abbreviation.xml"/gt

ltdesignation rdfparseType"Resource"gt
ltcontext rdfresource"CXT-Registry.xml"/gt
ltcontext rdfresource"CXT-Standard.xml"/gt
ltsign xmllang"en"gtMailing Address Country
Namelt/signgt lt/designationgt ltdefinition
rdfparseType"Resource"gt ltcontext
rdfresource"CXT-Legacy.xml"/gt ltcontext
rdfresource"CXT-Long Abbreviation.xml"/gt
ltcontext rdfresource"CXT-Medium
Abbreviation.xml"/gt ltcontext
rdfresource"CXT-Registry.xml"/gt ltcontext
rdfresource"CXT-Short Abbreviation.xml"/gt
ltcontext rdfresource"CXT-Standard.xml"/gt
lttext xmllang"en"gtThe name of the country where
the addressee is located.lt/textgt lt/definitiongt
lttype rdfresource"RCDIS.1.12116.1.xml"/gt
ltdomain rdfresource"VDALL.1.15147.1.xml"/gt
ltmeaning rdfresource"DCDIS.1.12800.1.xml"/gt
ltexample rdfdatatype"xsdstring"gtUnited
Stateslt/examplegt lt/DataElementgt
21
OWL, RDF XML Schema used to specify XMDR as UML
is used for 11179 metamodel
Types Cardinalities
Trang
Triples binary labeled relationships
What things go in own files? Which property
direction stored? Sequential ordering of
properties
22
XMDR Prototype Architecture Initial Implemented
Modules
External Interface
RegistryStore
Registry
Java
WritableRegistryStore
Subversion
RetrievalIndex
Jena, Xerces
LogicBasedIndex
FullTextIndex
Jena, OWI KS Racer,Kowari
Lucene
MappingEngine (defer)
Protege
Composition (tight ownership)
Generalization
Aggregation (loose ownership)
23
Inference
Disease
is-a
is-a
Infectious Disease
Chronic Disease
is-a
is-a
is-a
is-a
Heart disease
Polio
Smallpox
Diabetes
24
RDF Graph Query Facilities Compliment Text Query
Capabilities

• SQL-like queries
• e.g., names of ontologies in a registry
• Span items that are only indirectly connected
• e.g., data elements associated with a conceptual
  domain
• Expand queries to subsumed classes in hierarchy
• e.g., ConceptualDomain includes EnnumeratedConc..
• Transitivity
• e.g., all subclasses subsumed by a higher order
  class
• e.g., all superclasses (ancestors) of a
  particular class
• Least common ancestor
• e.g., closest subsuming concept for 2 concepts

25

• Deployment
• and Best Practices

26
Content Is King

• Proved true in Web 1.0
• Will remain true in Web 2.0, Semantic Web,
  Semantic Grids, etc.
• Ship early and often