Publishing to the Semantic Web Dr Owen Conlan Dr Alexander O’Connor

The World Wide Web HTTP HTML The Semantic Web Vision RDF Semantic Web Stack Web 2.0 / The Social Web Tagging Crowd Sourcing The Web of Data (Linked Data) VoID DBpedia Semantic Web Reasoning Logic, Rules Trust The Long Road for Semantic Web / ????

Tim Berners Lee driven “Linked Data uses a small slice of the technologies that make up the Semantic Web” Treat Schemas as Vocabularies Reuse existing schemas The Linked Data Movement

Community project with W3C support started in early 2007 Idea: take existing (open) data sets and make them available on the Web in RDF Interlink them with other data sets Linking Open Data Project

A Pretty (Scary) Diagram

DBpedia Transforming Wikipedia into a knowledge base Structure from Infoboxes HTML (titles) Categories Links other languages, redirects, disambiguations, etc Uses: as a controlled vocab, as an ontology Check out : ege,_Dublin ege,_Dublin (or Google “Trinity College Dublin dbpedia”)

Publish structured data in RDF on the web using URIs and shared vocabularies rather than the traditional Semantic Web focus on ontologies and inference Lowers barriers to entry Fosters widespread adoption Mature tools, techniques, patterns Linking Data

Formulated by Tim Berners-Lee (2006): 1.Use URIs as names for things 2.Use HTTP URIs so that people/apps can lookup these names 3.When someone/an app looks up a URI, provide useful information 4.Include links to other URIs so that they can discover more things This not an unambiguous specification, just a set of principles. Linked Data Principles

The Hypertext Transfer Protocol (HTTP) is an application-level protocol for distributed, collaborative, hypermedia information systems. It is a generic, stateless, protocol which can be used for many tasks beyond its use for hypertext, such as name servers and distributed object management systems, through extension of its request methods, error codes and headers. A feature of HTTP is the typing and negotiation of data representation, allowing systems to be built independently of the data being transferred. [RFC2616]

A Uniform Resource Identifier (URI) is a compact sequence of characters that identifies an abstract or physical resource [RFC3986] Syntax: URI = scheme “:” hier-part [“?” query] [“#” fragment] Example Note: scheme not the same as protocol URIs

Linked data needs dereferenceable URIs (ones we can use HTTP to retrieve a description of that resource) But we cannot serialise people, things over the internet (yet?) => we publish RDF documents on the web that describe them A real-word object != a document about that object e.g. creation-date for you != creation-date for your web- page Identifying Linked Data Resources

URI that identifies a real-word object != URI that identifies a document about that object Can make statements about object and can make statements about the document describing it How do we link these 2 URIs together? Identifying Linked Data Resources

303 Redirect (e.g. smith) Used for large, dynamic data sets Flexible because redirection can be separately configured for each resource e.g. can store data in multiple files or DB. Can change this at deployment/run-time. Typically used for resource descriptions in large data-sets URI styles for Linked Data

Fragment (e.g. smith) Used for small, static data sets Reduced number of HTTP round-trips => reduced latency A single HTTP request retrieves the entire document May transmit unnecessary data across the web Used for RDFa (defined via RDFa “about=” attribute) Typically used for vocabulary definitions URI styles for Linked Data

1.Create URIs for concept/thing and documents e.g. (URI identifying the person Dave Smith) (URI for RDF/XML document describingDave Smith) for HTML document describing Dave Smith) 2.Use HTTP redirects/content negotiation to access the desired resource description for the specific user agent 1.Client HTTP GET request on a URI identifying a object 2.Server recognizes URI, it answers using the HTTP 303 to send the URI of a description of the object 3.Client HTTP GET request on new URI 4.Server sends document from new URI 303 Redirect Approach

The picture below shows how dereferencing a HTTP URI identifying a non-information resource plays together with content negotiation: Simples… Huh?

1.Assign a URI to the RDF document defining the concepts e.g. (document URI) 2.Assign fragment identifiers to concepts within the document e.g Use HTTP requests to get the description 1.Client truncates a fragment URI to just refer to the document 2.Send HTTP GET to request the document 3.Server sends back the full document 4.Linked data application now inspects triples to find fragment Fragment Approach

Class <!--  Property Now we can refer to stuff

5 Steps to Publishing Linked Data

1.Understand the Principles 2.Understand your Data 3.Choose URIs for Things in your Data 4.Set up Your Infrastructure 5.Link to other Data Sets

A.Use URIs as names for things Anything, not just documents You are not your homepage Information resources (can be transmitted electronically) and non- information resources (cannot be transmitted electronically, e.g. a person!) B.Use HTTP URIs Globally unique names, distributed ownership Allows people to lookup those names Step 1: Understanding the Principles

C.Provide useful information in RDF when someone looks up a URI We can include RDF triple statements! D.Include RDF links to other URIs To enable discovery of related information e.g. via “follow your nose” browsing Relationship Links – to add context Identity Links – for URI aliases in other sources Vocabulary Links – to enable self-description Step 1: Understanding the Principles (cont.)

What are the key things in your data? People Places Events Book Flims Musician … This why domain expertise are critically important Step 2: Understand your Data

What vocabularies can be used to describe these? Principles: Reuse, don’t reinvent Mix liberally Examples: foaf -- Friend-of-a-Friend ontology geonames -- GeoNames ontology skos -- Simple Knowledge Organization System ckan.net Step 2: Understand your Data (cont.)

Step 2: Common Vocabularies bibo -- Bibilographic ontology cc -- Creative Commons ontology damltime -- Time Zone ontology doap -- Description of a Project ontology event -- Event ontolog foaf -- Friend-of-a-Friend ontology frbr -- Functional Requirements for Bibliographic Records geo -- Geo wgs84 ontology geonames -- GeoNames ontology mo -- Music Ontology opencyc -- OpenCyc knowledge base owl -- Web Ontology Language pim_contact -- PIM (personal information management) Contacts ontology po -- Programmes Ontology (BBC) rss -- Really Simple Syndicate (1.0) ontology sioc -- Socially Interlinked Online Communities ontolog sioc_types -- SIOC extension skos -- Simple Knowledge Organization System umbel -- Upper Mapping and Binding Exchange Layer ontology wordnet -- WordNet lexical ontology yandex_foaf -- FOAF (Friend-of-a- Friend) Yandex extension ontology

Use HTTP URIs Keep out of other people’s namespaces Create own URI and include alias information Abstract away from implementation details: Is better than this: bin/resource.php?id=/Berlin Use Natural Keys within URIs: Need to ensure the uniqueness of URIs Useful to base them on some existing primary key Whenever possible, use a key that is meaningful within the domain of the data set. e.g. use the ISBN as part of the URI of a book Step 3: Choosing URIs

Common patterns for URIs:  Thing  RDF  HTML Or use the file name extension: Step 3: Choosing URIs (cont.)

Describe the Data-set! e.g. dataset name, authorship, updates, licensing terms, crawler support, SPARQL endpoint location,... Vocabulary of Interlinked Datasets (VoID) A little later… Pick a Publication Pattern Is your input data: queryable, structured or text? What is the data volume? Is it static or dynamic? Test it Step 4: Set up Your Infrastructure

Step 4: Set up Your Infrastructure (cont.)

Popular predicates for linking owl:sameAs Foaf:depection Foaf:homepage Foaf:topic Foaf:based_near Foaf:maker/foaf:made Foaf:page Foaf:primaryTopic Rdfs:seeAlso Step 5: Linking

VoID (from "Vocabulary of Interlinked Datasets") is an RDF based schema to describe linked datasets A dataset is a collection of data, published and maintained by a single provider, available as RDF, and accessible, for example, through dereferenceable HTTP URIs or a SPARQL endpoint Step 5: Linking (cont.)

1.Understand the Principles 2.Understand your Data 3.Choose URIs for Things in your Data 4.Set up Your Infrastructure 5.Link to other Data Sets

Thank you!

1. 2.Debugging Semantic Web sites with cURL, sites-with- curl 3.Linked Data Tutorial, michael- hausenblas Linked Data Applications, M Hausenblas, DERI Technical Report Linked Data: Evolving the Web into a Global Data Space, Tom Heath, Christian Bizer References