TUTORIAL DAY - December 15th


SWAP 2008 features two half-day tutorials that include hands-on sessions. Tutorial are free for all SWAP 2008 registrants.
Please, when you register tick the tutorial(s) you plan to attend in order to make organization of classes more efficient.

NOTE: Participants are invited to bring their own laptops to be used during the hands-on sessions, which will be included in both the tutorial sessions.



Morning tutorial (9:30-13:00 CET)


Publishing Linked Data on the Semantic Web: how and why?
Danny Ayers (Talis)

Get the on-line material here

The Web is a global information space, already it has radically changed virtually all aspects of modern life.
It exists through the layering of a relatively simple protocol (HTTP) and text markup language (HTML) over the robust network that is the Internet.
The reason it is a unified space, and a key to its success, is the hyperlink - a simple means of leading the reader from one document to something related.
But the Web is predominantly document-based, even where the material presented is derived from a database, on the Web it generally appears in a human readable form.
The Semantic Web offers the promise of extending the capabilities of this system, with corresponding benefits to end users, into dealing with data first-hand. In short, we can do for data what we did for documents.
It is straightforward to translate from traditional means of representing data (typically RDBMSs) to representing the same on the Web, and many Semantic Web systems take the approach of creating a direct mirror of localised relational data on the Web. However, to take complete advantage of the Web infrastructure, we should look at it's glue, the hyperlink.
By considering the relations of databases in terms of relational links on the Web, it's possible to view the (Semantic) Web as a global database.
Global scalability can be provided by interlinked distribution of computer-friendly information, just like the Web. To maximise the benefits from their efforts, every Web developer should be aware of how and why to link data on the Web.

In the first half of this tutorial a review of basic (Semantic) Web principles will be presented, extending into description of issues that face data publishers.
The notions behind Linked Data will be presented, along with associated principles and practical techniques such as the publication of microformats/RDFa and intertwingling with existing open data.
Reasons why the typical Web developer should be aware of the benefits will be suggested, with a view to managing the inevitable evolution of the Web.

The second half of the tutorial will be hands-on, in which everyone will gain experience of linking data on the (Semantic) Web.


Afternoon tutorial (14:30-18:00 CET)


Ontology-based Data Integration
Giuseppe De Giacomo, Claudio Corona, Domenico Fabio Savo
(Università di Roma "La Spienza")

Get the on-line material here

Ontologies are concrete conceptualizations of a domain of interest. ŹNowadays, they are typically represented in terms of Description Logics, and are seen as the key technology used to describe the semantics of information at various sites.
The idea of using ontologies as a conceptual view over independent and autonomous data repositories is becoming more and more popular.
However, in order for this idea to become effective in practice, it is fundamental that the conceptual layer through which the underlying data layer is accessed does not introduce a significant overhead in dealing with the data. Based on this concerns, in the last years query answering over ontologies has been intensively investigated with a specific care to the computational complexity measured in the size of the underlying data.

This tutorial will present the main results of these studies and their impact on ontology-based data integration, including:

  1. The characterization of the tradeoff between expressive power of the ontology language, and efficiency in query processing;
  2. The DL-lite family of description logics, these are logics that have the nicest computational properties wrt accessing very large amount of data;
  3. Techniques for effectively delegating data processing to a relational DBMS or data federation tool, after a precomputation phase;
  4. Semantic mapping mechanisms to relate data at the sources with the ontology, which fully take into account the impedance mismatch between objects that represent instances of the ontology and values stored at relational sources.

The tutorial is divided in two parts. The first part will introduce the main concepts and techniques of interest in data integration through ontologies, both from an intuitive and formal point of view. The second part will consist of an hands-on session, in which the participants will use available automated tools for building an ontology and for connecting it to relational data sources through suitable semantic mappings.