Semantic Technologies

Over­view

Accord­ing to a 2009 Price­Wa­ter­house­Coopers report, “Semantic Web tech­no­lo­gies could revo­lu­tion­ize enter­prise decision mak­ing and inform­a­tion shar­ing”. By con­nect­ing more flex­ible, stand­ard­ized ways to model and share data with best prac­tices for identi­fy­ing the mean­ing (or, at the very least, the source) of descript­ive terms, Semantic Web tech­no­lo­gies open up new pos­sib­il­it­ies for devel­op­ing applic­a­tions that work across the web or behind your firewall.

In this course, we’ll learn about the build­ing blocks of the Semantic Web such as the RDF data model, the RDFa ver­sion that lets you embed machine-readable facts (or “triples”) into web pages, the SPARQL query lan­guage, and the Web Onto­logy Lan­guage (OWL) for defin­ing vocab­u­lar­ies and term rela­tion­ships. We’ll also learn about some of the open source and com­mer­cial soft­ware that lets you assemble these build­ing blocks into applic­a­tions that help you get more out of both your own data and the increas­ing amount of pub­licly avail­able linked data.

Classes for 2010

The Semantic Web: an Overview

Taught by Bob DuCh­arme.

The Semantic Web is a set of stand­ards and best prac­tices for shar­ing data and the semantics of that data over the web for use by applic­a­tions. What are the stand­ards? What are the best prac­tices? What does it mean to share semantics along with data, and how can that make the data more use­ful? How do applic­a­tions use data from across the web?

In this class, we’ll look at the high-level answers to these ques­tions, take a tour of the tech­no­logy and the acronyms, and see how they all fit together before the day’s remain­ing speak­ers dig deeper into the prac­tical use of these technologies.

Intro­duc­tion to Linked Data

Taught by Leigh Dodds.

Linked Data is a set of prin­ciples for pub­lish­ing data across the pub­lic web to max­im­ise its poten­tial for reuse. An increas­ingly large amount of data is being pub­lished using these guidelines, across a num­ber of dis­cip­lines and indus­tries, ran­ging from crowd-sourced data (e.g. dbpe­dia) and media out­lets (e.g. the BBC) to the US and UK gov­ern­ments. As momentum builds across the Linked Data approach to data pub­lish­ing, more pub­lish­ers and developers are explor­ing how to share their data in this way, and won­der­ing how to make use of the exist­ing resources to cre­ate new applic­a­tions or enrich exist­ing resources. In this class we’ll learn how to pub­lish Linked Data tak­ing a best prac­tices approach, as well as look­ing at how to con­sume linked data that has already been pub­lished. The les­sons will be illus­trated using the Talis Plat­form and some open source Linked Data browsers.

Lunch break, day one

 

SPARQL

Taught by Dr. Andy Seaborne.

SPARQL is the stand­ard W3C query lan­guage for semantic web applic­a­tions. It brings together the fea­tures of a num­ber of RDF query lan­guages into one method for extract­ing inform­a­tion from data rep­res­en­ted in RDF, whether small data­sets or large.

The next wave of SPARQL stand­ard­iz­a­tion is cur­rently under­way to add fea­tures that are use­ful for pub­lish­ing data and also to add mech­an­isms to update and man­age RDF data over the web.

This ses­sion will provide a solid ground­ing in SPARQL. After demon­strat­ing how power­ful some very simple SPARQL quer­ies can be, we will take a prac­tical approach to look­ing at the key fea­tures of SPARQL 1.0 and 1.1, and then explore the prin­ciples under­pin­ning the SPARQL query language.

RDF Mod­el­ing: Get­ting star­ted with RDFS, OWL, and SKOS

Taught by Leigh Dodds.

You’re start­ing to pub­lish Linked Data or assemble another RDF-based applic­a­tion, and you’ve ini­tially taken a pick and mix approach to select­ing from exist­ing vocab­u­lar­ies, but what hap­pens when you need to go bey­ond pop­u­lar vocab­u­lar­ies like FOAF, Dub­lin Core, or Goo­dRe­la­tions? How do you actu­ally go about mod­el­ling data using RDF tech­no­lo­gies? And how do semantic web schema lan­guages dif­fer from, say, XML schema?

This class will provide an over­view of RDF mod­el­ling, includ­ing the use of RDFS and OWL to cre­ate cus­tom vocab­u­lar­ies. The class will review how to make use of both tech­no­lo­gies and provide guid­ance for going deeper. It will also look at how vocab­u­lar­ies like SKOS fit into the pic­ture, and how it can used to help model a par­tic­u­lar domain.

End of day one

 

Put­ting Semantics on the Web with RDFa

Taught by Dr. Jeni Ten­nison.

This class will focus on the use of RDF that is embed­ded in XML, and espe­cially in HTML web pages, with the W3C’s RDFa stand­ard. We’ll look at how to cre­ate RDFa, how to add it to doc­u­ments manu­ally, and some strategies for auto­mat­ing the pro­cess. We’ll also see how to extract RDF triples from these doc­u­ments, and we’ll tour some of the large-scale web sites that are cur­rently mak­ing RDF data avail­able on the pub­lic linked data web using RDFa. Along the way, we’ll address these com­mon ques­tions about RDFa: how is it sim­ilar to micro­formats, and how is it dif­fer­ent? What kind of data is suit­able for expos­ing to to other applic­a­tions as RDFa? Is it only good for pub­lic web pages, or can it be use­ful behind a fire­wall? Can RDFS and OWL onto­lo­gies play a role in the use of RDFa?

SPARQL Update

Taught by Dr. Andy Seaborne.

In this follow-on class about SPARQL, we will intro­duce the new stand­ard fea­tures of SPARQL for update and man­age­ment of data using web pro­to­cols. SPARQL Update is a lan­guage for modi­fy­ing RDF data and SPARQL HTTP Update provides for REST­ful update of a col­lec­tion of RDF graphs.

Lunch break, day two

 

Applic­a­tion Devel­op­ment Workshop

The Applic­a­tion Devel­op­ment Work­shop will start off with a few slides to give del­eg­ates some ideas about the typ­ical struc­ture of semantic web applic­a­tions, but after that, it will be a dis­cus­sion between the fac­ulty for this entire course and the del­eg­ates. Bring your ideas for applic­a­tions and how you were think­ing of assem­bling them, and get the faculty’s opin­ions. How much cod­ing do you need to do? Do you need a spe­cial­ized triplestore? Should you use RDF/XML, n3, or RDFa? What are the open source options for tools, and what are the com­mer­cial ones? Which ones have been used where, and how did it go? The fac­ulty wants to hear your ideas, and they’ll be full of opin­ions about how to approach them.