1. RDF, RDFS making the Web a better place Ir. Richard Vdovjak, MTD. richardv@win.tue.nl

2. When the Keywords fall short Consider the following query: Give me pictures of Niagara Falls taken from the Falls Avenue in Niagara Falls, Canada with a tele-lens. the subject that is photographed the vantage point of the photographer the “tool” which is used

3. When the Keywords fall short This is what we are after:

4. When the Keywords fall short This is what we get:

5. When the Keywords fall short Or this:

6. So, what do we need to improve on that? A tool for specifying (rich) metadata (not just keywords). the metadata should include different taxonomies/ontologies of classes and resources that belong to those classes, e.g. a geographical ontology, ontology of lenses etc. properties within the classes of a certain ontology but also among different “neighboring” ontologies To insure scalability the rule “anything can say anything about anything” applies

7. Resource Description Framework (RDF) A foundation of processing metadata Issued by the W3C consortium www.w3.orgwww.w3.org While relying on XML for serialization, RDF focuses on semantics The formal semantics of RDF is defined see RDF Model Theory The aim is to make the Web “machine-understandable” not only machine readable an open-world framework that allows anyone to make simple assertions about anything it does not guarantee nor assumes consistency of these assertions distributed by its nature as is the WWW fully extensible

8. RDF in XML serialization “Ora Lassila” is the creator of the resource http://www.w3.org/Home/Lassila. Ora Lassila

9. RDF Data Model (A kind of) Directed Labeled Graph edges represent properties nodes represent resources (ellipses) literals (rectangles) are shared there may be multiple edges between two nodes there may be multiple edges with the same label pointing to different nodes the graph can contain cycles the graph is not necessarily connected subject object predicate

10. RDF Data Model as Triplets Definition: An RDF model M is a finite set of triplets (also called statements) of the form: M  R x U x (R  L) Where U is the set of (references to)resource identifiers or URIs L is the set of literals (string like elements) which denote the actual content (data) B is the set of blank nodes (nodes that don’t have a label) R= U  B is the set of all resources (both blank and labeled) subjectobject predicate

11. RDF Data Model as Triplets Example: [Pic1, filmFrame, “28”] [Pic1, depicting, Person1] [Pic1, createdBy, Person1] [Pic1, depicting, Person2] [Person1, age, “28”] 

12. RDF Data Model Implications: properties have always labels (URIs), i.e., there are no “blank” properties literal nodes cannot have properties (they can stand only as property values) it is possible to make statements about properties, but beware a statement about one property propagates to all properties with the same label in the model. In fact all properties with the same label are treated as one resource!

13. RDF Data Model as Triplets Example: [Pic1, filmFrame, “28”] [Pic1, depicting, Person1] [Pic1, createdBy, Person1] [Pic1, depicting, Person2] [Person1, age, “28”]  [depicting, how, “nicely”] depicting nicely how nicely how Any ideas?

14. RDF Data Model as Triplets Solution: [Pic1, depicting, b1] [b1, how, “nicely”] [b1, whom, Person1] [Pic1, createdBy, Person1] [Pic1, depicting, b2] [b2, whom, Person2] 

15. RDF built-in resource and properties rdf:Property a resource that represents (a type of) all properties rdf:type a property which says that a resources is an “instance” of another resource, e.g. [john, rdf:type, Person] Definition A model M is RDF-closed iff  [xxx,yyy,zzz]  M,  [yyy, rdf:type, rdf:Property]

16. Adding Gadgets: Containers rdf:Bag - a set with duplicates or an unordered list rdf:Seq - a sequence or an ordered list rdf:Alt -a list of “equivalent” alternatives rdf:li- a container membership property [container, lists, X] Committee rdf:Bag rdf:type John Jessie rdf:li Document approved Joseph rdf:li John Jessie approved Document Joseph approved The same?

17. Adding Gadgets Reification or statements about statements rdf:Statement rdf:subject rdf:object rdf:predicate Beware you can make assertions about the statement resource (i.e. not really about the property) Pic1Person1 depicting rdf:Statement rdf:subject rdf:object rdf:predicate rdf:type S nicely how

18. RDFSchema: adding more structure A modeling language on top of RDF (expressed in RDF) Introduces the following modeling primitives: rdf:Propety rdfs:Class rdf:typerdf:subClassOfrdf:subPropertyOf rdfs:Literal rdf:range rdf:domain rdfs:Resource rdfs:subClassOf rdf:type

19. RDFSchema: Classes resources may be divided into groups called classes Definition The extent of a class is the set of all resources that are linked to that class by the rdf:type property. two classes with the same extent are not necessarily the same, i.e. they might have different properties classes can be subsumed by other classes with the transitive rdfs:subClassOf property a class can be an instance of itself (!) this blurs the clean distinction of the schema and their instances known from databases in some cases it actually reflects the real-world, e.g. “Tulip” can be both a class (subclass of say “Flower”) and also its own instance when we are not interested in particularities of the instance, e.g. the picture depicts a tulip.

20. RDFSchema: Properties properties are linked to classes by specifying their rdfs:domain and rdfs:range,. i.e., they are defined externally wrt classes (they are first-class citizens) Definition The set D(R) of a property p is defined as the intersection of extents of all classes c indicated by [p,rdfs:domain (rdfs:range),c] Definition The extent of a property p is the set E of all triplets [x, p, y], where x  D and y  R If a property has more than one rdfs:domain (rdfs:range), objects(subjects) using this property are instances of all classes stated in the rdfs:domain (rdfs:range). properties can be (also) subsumed with the transitive rdfs:subPropertyOf property

21. RDFSchema:RDFS-Closure Definition A model M is RDFS-closed if the following holds: 1.  [x,y,z]  M,  [y, rdf:type, rdf:Property] 2.  [x,y,z], [y, rdfs:domain,u]  M,  [x, rdf:type, u] 3.  [x,y,z], [y, rdfs:range,u]  M,  [z, rdf:type, u] 4.  [x,y,z]  M,  [x, rdf:type, rdfs:Resource] 5.  [x,y,z]  M where z  U,  [z, rdf:type, rdfs:Resource] 6.  [x,rdfs:subPropertyOf,y] and [y,rdfs:subPropertyOf,z]  M  [x,rdfs:subPropertyOf,z] 7.  [x,y,z] and [y,rdfs:subPropertyOf,u]  M  [x,y z] 8.  [x, xrdf:type, rdfs:Class]  M  [x,rdfs:subClassOf, rdfs:Resource] 9.  [x,rdfs:subClassOf,y] and [y,rdfs:subClassOf,z]  M  [x,rdfs:subClassOf,z] 10.  [x,rdfs:subClassOf,y] and [a,rdf:type,x]  M  [a,rdf:type,y] These can be also considered as inference rules, i.e. how to derive an RDFS-closed graph

22. RDFSchema: Example

23. RDF(S) Applications (Web) metadata CC/PP Composite Capabilities/Preferences Profiles P3P Platform for Privacy Preferences Data integration (programmable mediators) foundation for higher ontology languages OIL, OWL RDF(S) query languages Web needs not only the metadata but also a means to reason about them. RQL, RDQL, etc.

24. RDF(S) Pitfalls Unrefined Datatypes XML data type system can be adopted Reification needs an external interpretation it is application dependent, i.e. there might be differences in interpretations More expressive power? price to pay in terms of performance, scalability sometimes not needed There are not so many RDF metadata available on the Web this is not a shortcoming of RDF(S) triplets can be harvested from X-link but mainly generated from the DB backends

25. RDF(S) Pitfalls 2, The Designation Problem Correspondence between URIs and real-world objects how do we assure that two (distributed) sites when referring to the same real-world object are using the same URI? if they don’t how do we reconcile (not express) two different URIs pointing to one object? Given reconciliation? By whom? a “Web URI Institution”? Value based decision? two URIs having the same value do not have to be the same real- life objects... Source 1 Source 2 John1234 John

26. The End of Part 1 Let’s have a break!