Interaction Design for the Semantic Web
User System Interaction

This module is led by Lynda Hardman, and runs from 10 until 14 November 2008.

The Problem

Large amounts of information are available on the Web. More and more information, however, is becoming available on the Semantic Web.
How can we use the explicit relations encoded in this content to improve the user interaction with the information?
How can we make use of these relations to present the information to the user in an appropriate way?
What are the enabling Semantic Web technologies, what do they enable and what help can they provide?

Table of contents

Schedule

Mon 10 10:00 Goals of module, course introduction (see also inaugural speech).
Semantic Web introduction.
Start assignment 1.
Tue 11 Complete assignment 1.
09:30 Wake up video before Semantic search. Article giving an overview of the semantic search survey: An analysis of search-based user interaction on the Semantic Web.
Facet browsing in /facet - watch the movie and follow the slides.
mSpace, CHIP and /facet (click on one of the headings All Objects, Style or Artists).
Parallax video - exploring freebase data.
Alia Amin: Understanding Information Seeking Needs
Start assignment 2.
Wed 12 Work on assignment 2.
Thu 13 Complete assignment 2.
10:00 Michiel Hildebrand: Interactive exploration of heterogeneous cultural heritage collections and Design and evaluation of an open annotation tool.
Meaningful presentations, Vox Populi talk and demo.
Bringing the IPTC News Architecture into the Semantic Web.
Start group work on assignment 3.
Fri 14 Continue work on assignment 3.
13:30 Present work on assignment 3 to group.
Finalise work on and write report for assignment 3.

Assignments

Assignment 1 (4 hours)
Individual assignment

The goal of the assignment is to explore the relationship between potential questions users may have and how they can find the information from existing web sites. The example domain is cultural heritage collections.

Explore the provided example cultural heritage web sites and write an analysis (a brief description of what you are discussing, along with the advantages and the disadvantages) of the following:

Use the following sites for your analysis:
the Rijksmuseum collection,
Louvre,
The Artchive then click on the Mona Lisa image.

Think of some questions you would like the answer to from each of the sites. Write about your experiences in finding the answer. Try using search facilities on the site. Try exploring the site through navigation.
Given the information available from these three sites, how would you want to access all the information?
Write down one or two questions whose answer requires information from at least two of the sites.
Write down an example of extra information you would want that is not available from these three sites.

Produce your report in HTML of approximately 2 pages (max. 1000 words). Illustrations are desirable, and can be linked to or included. Publish the report on your web page and email the URL to Lynda.Hardman@cwi.nl before 10:00am Tuesday 11 November 2008.

Assignment 2 (6 hours)
Individual assignment

The goal of this assignment is to explore experimental interfaces facilitated by underlying ontologies.

CHIP, Eculture demo.

For all displays to work for the Eculture demo you should install Firefox 2.0 or later (as described at the beginning of the tutorial).
The relation search in the Eculture project is a prototype which demonstrates how two entities (artefacts, people, places, concepts, etc.) can be related with each other. A user can specify two items of interest, e.g. two artefacts, or an artefact and a person. The system then provides several relationships as a result. See the tutorial for further explanation. If you encounter any bugs in the system then we would be grateful for a description of them. Please send them to Alia.Amin@cwi.nl.

Think of some questions you would like the answer to. Use the different facilities from the different systems and provide comments on which were helpful and how.
Write down two questions whose answer requires linking more than two pieces of information.
Comment on the very different interfaces of these sites.

How would you improve these interfaces? Is there other functionality you would like to see added?

Produce the report in HTML of approximately 3 pages (max. 1500 words). Illustrations are desirable, and can be linked to or included. Publish the report on your web page and email the URL to Lynda.Hardman@cwi.nl before 10:00 Thursday 13 November 2008.

Assignment 3 (1 day)
Group assignment

You are the representative of a start-up company that develops semantic web technology. You are asked to work on a design for a client by the company. Select one of the following scenarios, or create your own. (Check with me before starting on the application design stage, and before spending too much time on it.)
While the preferred domains are artworks/cultural heritage or news, these should not be seen as restrictive.

Museum scenario
A museum would like to make an exhibition appealing to members of the general public from artefacts in its own and sister collections. The organizers need to browse the collections, find artefacts belonging to a number of themes and build a story around them. Design an application (functionality and example key interfaces) to support the task of the exhibition organizers in selecting and collecting artefacts, creating a theme, building a storyline and relating the artefacts to the storyline.

News scenario
A news agency wants to provide personalised access to multiple national and international news feeds to allow subscribers to:

Holiday scenario
A family wants to plan their own holiday, where they can choose different places to stay, be able to travel between them, have access to activities that appeal to the whole family. Keep track of topics such as place, hotel, restaurants, activities. (Work on this topic is taking place in the EIFFEL project (in French), of which Mondeca is one of the partners.)

On the basis of your chosen scenario create a user-centric application design. Decide on the desired functionality. Explain the required underlying domain and/or discourse semantics. Be aware that while the semantic relations are necessary for supplying the functionality, the user does not need to be aware, and perhaps should not be aware, of them. How much explicit semantics to expose to the end-user is one of the design decisions you should document.

Part of the application design should contain at least one of the following, as appropriate:
interface for selecting and comparing items, be they people, places or objects;
interface for gathering and browsing "interesting" items;
interface for inputting user preferences into system (e.g. artwork style, favourite artist, countries of interest, politicians, hotel facilities, dietary requirements).

You should see the report as a means of conveying your design to a client. In the report, explain the reasoning behind the design as folows.

Publish your assignments on your web page and email a single URL to Lynda.Hardman@cwi.nl before 09:00 Monday 17 November 2008. Make sure you include the names of the people in your group in the email and in the report.

Web Sites used in the course

The Rijksmuseum collection, Louvre and The Artchive are cultural heritage collections.
mSpace, CHIP and /facet allow exploratory access to (cultural heritage) information.
Semantic search survey by Michiel Hildebrand.
These show interesting interfaces:

These are more oriented to search.

Vox Populi, a system for generating video sequences based on argument structures.

Semantic Web Resources

How Will We Interact with the Web of Data? by Tom Heath.
Semantic Web Challenge, in particular the 2008 winner paggr.

Wordnet is an existing English language thesaurus from Princeton that has been converted to RDF.
Dublin Core (DC), VRA Core are vocabularies for describing resources. VRA Core is specialised for works of visual culture as well as the images that document them.
SKOS, Simple Knowledge Organisation System, is a family of formal languages, built on RDF and RDFS, designed for the representation of structured controlled vocabularies, such as thesauri or classification schemes.
FOAF (Friend of a Friend) a means of creating links among people on the Web, also the FOAF project, in particular the explorer.

Random Useful Links

These are some links that may not be useful now for the course, but are useful resources about relevant topics.

Web Design

Semantic Web

Facet Browsing


Valid XHTML 1.0 Strict