Declarative Languages for Querying Portal Catalogs

Christophides Vassilis December 2000

ICS-FORTH & Univ. of Crete

Declarative Languages for Querying Portal Catalogs V. Christophides D. Plexousakis G. Karvounarakis S. Alexaki Computer Science Department, University of Crete Institute for Computer Science - FORTH Heraklion, Crete 1

Christophides Vassilis December 2000

ICS-FORTH & Univ. of Crete

What is a Community Web ? n

A group of people sharing a domain of discourse and a set of information resources (e.g., data, documents, services) and having some common interests g

n

Commerce, Education, Health

The main requirement is to provide a single point of useful, ubiquitous comprehensive, and integrated access to various information resources g

Web Portals

Workplace

Education

C-Web

Commerce

Health

2

1 1

Christophides Vassilis December 2000

ICS-FORTH & Univ. of Crete

Portals Classification

Existing Communities

On-line Communities

3

Christophides Vassilis December 2000

ICS-FORTH & Univ. of Crete

Portals Classification

4

2 2

ICS-FORTH & Univ. of Crete

Christophides Vassilis December 2000

Knowledge Processes in Corporate Communities

Generating new knowledge

Representing knowledge in documents and databases

Dissemination of knowledge within organisation

ICS-FORTH & Univ. of Crete

Accessing knowledge from external sources

Embedding knowledge in services and processes

Using knowledge in decision making 5

Christophides Vassilis December 2000

Community Web Portals: Main Features Advanced Knowledge Schemas (ontologies, thesauri)

Heterogeneous resource descriptions

Complexity and diversity of information resources

6

3 3

Christophides Vassilis December 2000

ICS-FORTH & Univ. of Crete

What we need? n

Flexible Modeling of Community Web Portal Catalogs u Exploit existing forms of community knowledge • Ranging from simple vocabularies to formal ontologies

v Describe in various ways community information resources • Administration, Classification, Content Rating, Channels, …. n

Secondary Storage Management of Portal Metadata u Large Schemas: e.g., 100 Mbytes of ODP Topics (the Art Hierarchy contains 20000 terms)

n

v Voluminous Description Bases: e.g., 700 Mbytes of ODP indexed URLs Declarative Query Languages for Portal Catalogs u Interleave schema with data querying v Optimize access to Portal Catalogs 7

Christophides Vassilis December 2000

ICS-FORTH & Univ. of Crete

Our Approach RDF Virtual XML Warehouse

High-level Access to

Documents

Archives

community information Web Databases

Å

Use W3C Standards to describe (RDF/S) & exchange (XML) information

Å

Our Contribution: Declarative Languages for Browsing & Querying

8

4 4

Christophides Vassilis December 2000

ICS-FORTH & Univ. of Crete

Outline n

Example of a Portal Catalog for Cultural Communities

n

Describing and Querying Community Resources g A Formal Data Model for RDF/S g The RDF Query Language (RQL)

n

Architecture g Core Middleware (RDF Store, Parser/Loader, Query Interpreter)

n

Build Community Web Portals g Create/Extract Resource Metadata g Develop Specific Applications for Communities g Optimize RDF Storage and Queries

n

Summary 9

Christophides Vassilis December 2000

ICS-FORTH & Univ. of Crete

Building a Cultural Community Web Portal using RDF String

fname

String

lname

Artist

exhibited

Artifact

Painter

“Rodin”

Sculpture

paints

Painting

lname

Portal Resource “Picasso” Descriptions “Pablo”

r2: museoreinasofia.mcu.es / guernica.jpg

last_modified

technique

String title

ExtResource

String

creates &r1

&r5

Web Resources

Museum Date

sculpts

Sculptor

Portal Schema

creates

exhibited paints

fname lname &r6

&r2

paints &r3

r1:www.rodin.fr / thinker.gif

last_modified &r4 title

technique last_modified

“oil on canvas”

2000/06/09 “Reina Sofia Museum”

2000/01/02

r3:www.artchive .com/ woman.jpg

r4:museoreinasofia.mcu.es

10

5 5

ICS-FORTH & Univ. of Crete

Christophides Vassilis December 2000

Resource Description Framework (RDF/S) n

n

RDF: Resource Descriptions g Data Model: Directed Labeled Graphs • Nodes: Resources (URIs) or Litterals • Edges: Properties – Attributes or Relationships • Labels: Nodes (Class names) and Edges (Property names) • Statement: assertion of the form resource, property, value • Description: collection of statements concerning a resource g XML syntax RDF Schema (RDFS): Schema Vocabularies g Specialization of both classes & properties (simple & multiple) g Multiple classification under several classes g Unordered, optional, and multi-valued properties g Domain and range polymorphism of properties

ICS-FORTH & Univ. of Crete

11

Christophides Vassilis December 2000

RDF/XML Serialization: Data ….

Reina Sophia Museum 2000/06/09

ICS-FORTH & Univ. of Crete

…. 13

Christophides Vassilis December 2000

RDF/S vs. Well-Known Formalisms n

Relational or Object Database Models (ODMG, SQL) g Classes don’t define table or object types g Instances may have associated quite different properties g Collections with heterogeneous members

n

Semistructured or XML Data Models ( Ï Å Ì, UnQL, YAT, XML Schema) g Labels only on nodes or edges g Class and property subsumption is not captured g Heterogeneous structures reminiscent to SGML exceptions

n

KR Languages (Telos, DL, F-Logic) g Absence of complex values and n-ary relationships (bags, sequences) g Inadequate reasoning support 14

7 7

ICS-FORTH & Univ. of Crete

Christophides Vassilis December 2000

A Formal Data Model for RDF n

Ç(Í, c < Class

gp

∈ P => p < Property

g p1,p2 n

∈ P and p1 < p2 => domain(p1) ≤ domain(p2 ) and range(p1) ≤ range(p2)

Type System: τ = τL | τU | {τ} | [τ] | (1:τ + 2:τ + … + n:τ)

n

Interpretation Function: g Literal

types, [[ τL ]] = dom(τ L)

g Bag

types, [[ {τ} ]] = {ν1, ν2,…, νn}, ν1, ν2,…, νn ∈ V are values of type τ

g Seq

types, [[ [τ] ]] = {ν1, ν2,…, νn}, ν1, ν2,…, νn ∈ V are values of type τ

g Alt

types, [[ (1:τ 1 + 2:τ2 +…+ n:τn ) ]] = νI , νi ∈ V, 1