Architecture Patterns for Semantic Web Applications
- 1. NONOSQLArchitecture Patterns for Semantic Web ApplicationsBrian PanullaPenn State Web 2011 Conference
- 2. Twitter: @bpanulla(c)2010 Google, Imagery (c)2010 TerraMetrics, NASA, Map data (c)2010 Europa Technologies, Google, INEGI, AND
- 3. Three things the Semantic Web is not(mostly)
- 7. Who cares?
- 9. On NoSQL...
- 11. NoSQL: Key/Value and Graphs
- 14. When things go wrong
- 15. Linked Data - 2008http://richard.cyganiak.de/2007/10/lod/
- 16. Linked Data - 2008http://richard.cyganiak.de/2007/10/lod/ Last updated: 2010-09-22
- 18. RDF: Framework for knowledge representationDeclares resourcesSpecifies properties of resourcesW3C standard since 1999http://www.w3.org/RDF/
- 19. What is a Resource?Everything on the web (or not) is a resource.UniformResourceLocator
- 20. RDF is not XML...but occasionally dresses up like itXMLN-TriplesN3TurtleJSONRDFa in HTML...
- 21. URIsResources are identified by a Uniform Resource Identifier (URI)Can be found in your XHTML DTD and html tag’s xmlns attribute:
- 22. URIs... Are not necessarily URLshttp://webconference.psu.eduhttp://brian.panul.lahttp://alumni.psu.edu/brian.panulla
- 23. TriplesRDF is expressed as triples of URIs:subject (“Penn State”) - resourcepredicate (”is a") - propertyobject (”University”) - resourceIs aPennStateUniversity
- 24. TriplesAnother example:subject (“Brian Panulla”)predicate (”presented at")object (”psuweb11”)presented AtBrianPanullapsuweb11
- 25. Triples to GraphsUnitedStates ofAmericaCity ofPortlandState ofOregonNorthAmerica State ofWashington State ofCaliforniaCanadaNATOIs abordersIs inHeld AtIs inMember ofpresented AtIs inbordersMember ofattendedIs inIs inIs aIs located inIs inPSUWEB11PennStateBrianPanullaPortlandStateUniversity
- 26. RDF SchemasRDFS provides limited Set Theory featuressubClassOfsubPropertyOfDomainRangeThink of classes in RDFS as sets rather than OOP classes
- 27. OntologiesOntology is the study of being or reality.A Formal Ontology is a specification of a conceptualization (Gruber, 1995)Aristotle
- 28. Defining an ontologyWeb Ontology Language (OWL)Version 2.0 (October 2009)http://www.w3.org/TR/owl2-overview/Yes, rly.
- 29. Some RDFS/OWL FeaturesClassesSub-classEquivalent ClassesDisjoint ClassesCardinality constraints (max/min)IndividualsSame IndividualPropertiesSub-propertyEquivalentInverseSymmetricTransitive
- 31. Triple StoresGraph database that knows RDFQuad Stores: triple store that includes provenance “where did the data come from?”Various stores provideTransactional / non-TransactionalIn-Memory / File system / RDBMS storage
- 32. Popular Triple StoresHP/Apache JenaPellet ReasonerStorageIn memoryRDBMS (transactions)Filesystem storage (high performance)Open Source
- 33. Popular Triple StoresFranz AllegroGraph (server)Virtuoso Universal Server (server)Mulgara (RESTful service, OSS)Redland (C, Obj-C)NewStardog - Integrated reasoning databaseRDFa API - https://github.com/webr3/rdf.js
- 35. Del.icio.us
- 36. OMNOMino.us
- 37. Basic RequirementsInformationPeopleBookmarksTagsFunctionalityCreate an accountAdd a bookmarkDelete a bookmarkBrowse other users’ bookmarks
- 38. Logical Schema
- 39. Relational DB - Physical Schema
- 40. Modeling in RDFMinimize artificial entitiesRe-use existing schemas or ontologies
- 41. Useful SchemasDublin CORE METADATA INITIATIVE (DCMI)FrienD-of-a-friend(FOAF)
- 42. DMCI
- 43. FOAFWell-defined ontology for people, organizations, and social networks
- 44. FOAF
- 45. RDF Data Model - Memberfoaf:Personfoaf:fullnamefoaf:nicknamefoaf:mboxfoaf:openidLiteral (“datataype”)propertiesResource (“object”)property
- 46. RDF Data Model - Bookmarkfoaf:Documentdct:titledct:descriptiondct:createddct:modifiedfoaf:pagefoaf:topicLiteral (“datataype”)propertiesResource (“object”)properties
- 48. Custom properties and classesomnom:bookmarkedhttp://omnomino.us/ontology/om.n3#bookmarkedomnom:bookmarkedfoaf:Documentfoaf:Person
- 49. Custom properties and classesrdfs:subClassOfrdfs:subClassOffoaf:Documentfoaf:Personomnom:Memberomnom:Resource
- 50. RDF Data Model
- 51. SPARQLGraph query languageRead-only (1.0)Safe to expose endpoint to WebSPARQL 1.1 adds updates
- 52. A Simple SPARQL ExampleSELECT ?subj ?pred ?objWHERE { ?subj ?pred ?obj.}Demo
- 53. SPARQL Query TypesSELECT – Returns tuples matching specified patternASK – yes or no (tests for existence)CONSTRUCT – Returns a graphDESCRIBE – Returns a graph determined by the query engine
- 54. SPARQL Examples: ASKASK { ?urifoaf:openid <http://brian.panul.la>.}
- 55. SPARQL Examples: MemberSELECT DISTINCT ?uri ?fullname ?nickname ?emailWHERE { ?uri a omnom:Member. OPTIONAL {?urifoaf:name ?fullname;foaf:nick ?nickname;foaf:mbox ?email. }FILTER (?uri = <http://omnomino.us/member/24601>)
- 57. Updates via API
- 58. The Good:Domains where schemas/models change rapidly or data is sparseSemantics of relational model inadequate (e.g. inferencing, inheritance)Domains emphasizing relationshipsSocial networksTaxonomies
- 59. The Bad:WeaknessesBad for opaque objects with few relationshpsLarge sets of homogenous objectsAlternativesRDBMSNoSQLDocument DBsKey/Value StoresGraph databases
- 60. Linked DataPublish/Syndicate complete information setsEmbedded explicit semantics, unique identifiersHave minimal impact to other Web information publishingMay be static or dynamically generated
- 61. ResourcesSemantic Web Programming - John Hebeler, Matthew Fisher, Ryan Blace, and Andrew Perez-LopezSemantic Web for the Working Ontologist: Effective Modeling in RDFS and OWL -Dean Allemang and James HendlerProgramming the Semantic Web by Toby Segaran, Colin Evans, and Jamie Taylor
Editor's Notes
- #2: http://www.flickr.com/photos/mayhem/3104825022/in/photostream/
- #12: http://www.flickr.com/photos/herr_hartmann/448478567/in/photostream/http://www.flickr.com/photos/mayhem/3104825022/in/photostream/
- #29: http://www.flickr.com/photos/jingleslenobel/4971486823