When You Can’t Read It, Listen to It! An Audio-Visual Interface for Book Reading

C. Duarte and L. Carriço, When You Can’t Read It, Listen to It! An Audio-Visual Interface for Book Reading, vol. 5616 of Lecture Notes in Computer Science, pp. 24–33. Springer Berlin / Heidelberg, 2009. [PDF]


This paper describe the Digital Talking Book player for mobile devices, an application that was designed to allow people with visual impairments to learn from audiobooks. It allows the user to replicate the same functionalities of a paper book on digital content. The paper presents an usability evaluation of the prototype.


The book is talking to you – using an audio version of the course textbooks to support learning

S. Zerachovitz and M. Zuker, “The book is talking to you – using an audio version of the course textbooks to support learning,” in Proceeding of ITHET07 conference, (Kumamoto, Japan), June 10-13 2007. [PDF]


This paper presents subjective-reported data that support the thesis that audio-books of course material is helpful for students with learning disabilitites and for those whose main language is not the one the course is taught. The authors found that students like tghe convenience of learning whenever and wherever they choose. Many student listened to the audio books while reading the printed book. Learning by the audio book si more passive than learning by reading. Many students felt that learning while listening to the audio book increased their comprehension.

Spoken language technologies applied to digital talking books

I. Trancoso, C. Duarte, A. Serralheiro, D. Caseiro, L. Carriço, and C. Viana, “Spoken language technologies applied to digital talking books,” in Proceedings of Interspeech, (Pittsburgh, PA, USA), September 17-21 2006. [PDF]


This paper presents the DTB player, which offer to visually impaired users an evolution of paper books. The prototype offers a multimodal interface that presents the textual content of the book synchronized with an audio narration, either pre-recorded by a human speaker or constructed using a text-to-speech synthetizer. Speech recognition allows further the user to add bookmarks or annotation to the book. This paper summarized the different language technologies that may be integrated in spoken books and the different application domains in which spoken books might be used.


Draquila: Italy trembles

Today I am happy: Draquila, the movie, was presented at Cannes festival: http://bit.ly/bKBTIg looking forward to seeing it.

Why do Italians vote Berlusconi? The violence of propaganda, the impotence of citizens, questions of the economy, illicit power relationships…

And a catastrophe: the city of L’Aquila devastated by an earthquake… all these combine to show how the young Italian democracy has been subdued. The caricature of Berlusconi – one of the director’s most celebrated impersonations – strolls through Aquila’s refugee camp and wanders the deserted town like an emperor at the end of his reign. A town devastated by an earthquake – the perfect location from which to recount Italy’s drift into authoritarianism, the mess of blackmail, scandal, swindles and inertia of the political classes, the media and the citizens, that have paralysed the country. Why do the Italians vote for Berlusconi? Why do they consider democracy an unsuitable system of government? Aquila – this magnificent city laid low by an earthquake – will give us the answers. Why did the proud people of Aquila exchange their most precious commodity – their community, a dynamic town full of students and works of art – for a little apartment in a dormitory town, furnished by Berlusconi? Why did they believe TV propaganda rather than the evidence of their own eyes? And how did it happen to others too, as quickly and as deceitfully? Who was leaning on them? The days of Berlusconi’s reign seem numbered: it’s time to search through the rubble and draw what conclusions we can. —Cannes Film Festival, 2010


RT @karenchurch : Summary of WWW 2010

My colleague Karen Church wrote a great report of WWW2010:

As usual with multiple tracks and loads of great talks it was difficult to choose between sessions at this years WWW, in Raleigh, North Carolina. I focused my attention on the keynotes, panels and technical sessions related to interfaces, users profiling, and search.

(1) Mobile will be big: although there was no specific mobile track at this years WWW and although there was very few mobile-related papers (ours was one of 2 in the entire conference!!!) – one of the key trends mentioned in panels, future web sessions and in all of the key notes is the future of the mobile web and the importance of mobile handsets as pervasive information access devices. Vint Cerf’s keynote pointed to the fact that only 25% of the worlds population access/use the internet through desktops which according to him means he still needs to “convert” 75% of the worlds population! Vint pointed to the fact that there are almost 5 billion mobile users worldwide, and for many their mobile handsets will be their first point of contact to the mobile Internet, thus making it possible to reach higher levels of mobile internet penetration. It appears that mobile will be a bigger trend at next years conference

(2) The future of search according to people from Yahoo, Google and Microsoft is (a) getting to the long tail, (b) intelligent facets and improve interfaces, (c) moving to mobile search and (4) social search. I attended a very interesting panel called “Search is dead: long live search”. The panelists were Marti Hearst (she wrote a very nice book on search user interfaces which is available free to download online: http://searchuserinterfaces.com), Barney Pell from Bing/Microsoft who is big into voice-enabled mobile search, Andrew Tomkins who is director of engineering at Google and Andrei Broder from Yahoo! Research. Prabhakar Raghavan the Head of Yahoo! Labs acted as moderator. You can see the whole thing via video here: http://qik.com/video/6360405

(3) Twitter, twitter, twitter: there were lots of twitter-related papers at this years WWW and in the Web Science (WebSci) conference being co-held in the same venue. This blog article summarizes all the twitter related papers: http://blog.marcua.net/post/566480920/twitter-papers-at-the-www-2010-conference

The take away message is that all the big players are trying to incorporate social networks into the online search experience. Search is not going to be an isolated activity any more.

(4) I attended some tutorials related to web search behaviour. The most relevant/interesting was “Recent Progress on Inferring Web Searcher Intent” given by Eugene Agichtein, from Emory State. He presented lots of detail on how we can try to gather information regarding the intent of web users in their search tasks through log analysis, click-through behaviour, eye-tracking and mouse movements, etc. Interesting/relevant for anyone working in search, web behaviour or user profiling.

This reminds me of Anne Aula’s talk at CHI2010 where she presented a similar piece of research that was demonstrating that doing some machine learning on top of repeated search strings it was possible to infer whether the user was getting frustrated.

(5) A new track this year called Future Web involved discussions/interviews with various leaders in the field on future trends on the web related to politics, environment, social, mobile, etc. They have a YouTube channel so you may be able to catch up on some of these interviews: http://futureweb2010.wordpress.com/ and http://www.youtube.com/user/Futureweb2010#p/u

(6) Best papers award was given to a Recommender Systems paper: All recommender systems related papers were mainly in the personalization track http://www2010.org/www/2010/04/best-paper-awards/

(7) Lots of advertising – those interested check out the internet monetization tracks

(8) Social networks – 3 tracks dedicated to this pretty hot topic – again you can check the papers online

(9) Danah Boyd keynote focused on big data and privacy issues – she challenged the audience with the following – just because you have access to lots of data does not mean you should work with it. According to Danah we should all be more concerned with ethical questions associated with the data/the users rather than than access to data itself. Here’s a summary of the talk: http://www.danah.org/papers/talks/2010/WWW2010.html

This goes against the Linked data and semantic web movement. Their motto is get the data out there and then we will figure out what to do with it.

Some nice papers related to search that are worth a mention:

(1) Kumar & Tomkins, “A Characterization of Online Search Behaviour”: The authors look at online search behaviour using a dataset from the Yahoo search and toolbar logs. The dataset is over a year old at this point and as such some characteristics may have changed. The authors propose a new taxonomy of pageviews. The paper shows that1/3 of page views are for content, approx. 1/3 are related to communications while approx. 1/6 are search, however, the authors go on to show that although explicit search activity is low, this activity leads to increased browsing/content accesses by users.

(2) Horowitz & Kamvar: “The Anatomy of a Large-Scale Social Search Engine”. Following on from the original Google paper presented by Brin & Page in WWW 1998, the Aardvark team (who now belong to Google) provided an overview of their social search engine at this years WWW. A very nice read, describing the search engine architecture/algorithms used and an overview of the behaviour of its users.

Note: All papers are available online

Next years WWW will be in India: and the deadline for papers is always at the start of November! http://www2010.org/www/2010/05/www2011/

Thanks Karen for sharing these notes.

Dutch city launches iPhone app for lodging civic complaints

This reminds me of FixMyStreet a project I blogged about three years ago. As my fellow readers know I really much like these sorts of application to engage citizen in a more active engagement with local government.

Potholes, stray garbage, broken street lamps? Citizens of Eindhoven can now report local issues by iPhone, using the BuitenBeter app that was launched today. After spotting something that needs to be fixed, residents can use the app to take a picture, select an appropriate category and send their complaint directly through to the city council. A combination of GPS and maps lets users pinpoint the exact location of the problem, providing city workers with all the information they need to identify and resolve the problem.

Website: www.buitenbeter.nl

Related projects:

In San Francisco, civic complaints via Twitter

NYC challenges developers to create apps using city data

Tagging repairs for local government


Location-based services for mobile telephony: a study of users’ privacy concerns

L. Barkuus and A. Dey, “Location-based services for mobile telephony: a study of users’ privacy concerns,” in Prooceedings of the INTERACT’03, 2003. [PDF]


This paper present an interesting study of privacy concerns for location-based services. The author used a diary approach. They asked 16 participants to generate a 5-day journal in which they answer pre-specified questions about the usefulness and level of concern in using presented location-based services. A subset of participant was interviewed after completing the the journal for completing their entries. Also, the other interesting methodological approach that the author used was that of asking participants to imagine the existance of LBS services that would use or display their location.


Ethnographic methods to study context: An illustration

M. Evans, D. Leake, D. F. Mcmullen, and S. Bogaerts, “Ethnographic methods to study context: An illustration,” tech. rep., Pervasive Technology Lab, Indiana University, 2005. [PDF]


This paper describe a case study where ethnographic research methods are applied to understand contextual factors that play a role for distributed collaborative troubleshooting. The authors conducted a nine-month naturalistic study of real-world remote diagnosis of electronic devices by ad-hoc teams.

Ethnographic studies can be conducted with a number of strategies, including controlled and quasi-experiments, surveys, histories, archival analyses, and case studies. According to Yin (Yin 1994, pp. 1–9), the unique advantages of each depends on three conditions: (a) the form of the research questions(s); (b) the control over actual behavioral events; and (c) the focus on contemporary as opposed to historical phenomena.

A diary study of mobile information needs

T. Sohn, K. A. Li, W. G. Griswold, and J. D. Hollan, “A diary study of mobile information needs,” in CHI ’08: Proceeding of the twenty-sixth annual SIGCHI conference on Human factors in computing systems, (New York, NY, USA), pp. 433–442, ACM, 2008. [PDF]


This paper present a user study of how and why information needs arise when the user is on the go. The study reports a diary study of 20 people over the course of two weeks. They examined the information needs the participants had and the strategies that they used to address these needs.They also focused on the contextual factors thar prompted each need and influenced how it was addressed.

The authors used a snipped technique which consists in sending a short SMS to capture the gist of the moment and a web diary to provide more structured information around that moment. At the end of the day participants logged into the website to answer six questions about their snippet:

1. Where were you?

2. What were you doing?

3. What was your information need?

4. I addressed the need (At the time, Later, Not at all)

5. If you attempted to address the need, how did you do so? If you didn’t make an attempt, why didn’t you?

6. Could you have addressed your need by looking at your personal data (e.g., email, calendar, web browsing history, chat history, or other)

Through the 421 generated entries they were able to define the following taxonomy: trivia needs (18.5%), prompted by conversations or location-specific artifacts; directions (13.3%); friend info (7.6%); business hours, phone numbers (7.1%); personal schedule (6.4%); movie times (2.4%); and travel related.

Participants indicated that 72% of their reported information needs were prompted by some contextual factor. The contextual prompting can be classified in four broad categories: Activity, Location, Time, and Conversation. Activities reflect what the person was doing at the time. Location is the place where the person was at and includes any additional artifacts at that specific location. Time is the time when the need arose, and conversation is any phone or in-person conversation the participant was involved in at the time. Some diary entries were related to multiple aspects of context, such as having a conversation with someone about artifacts at the current location.


Figure 5. Percentage of different contextual factors that prompted information needs

The study reports qualitative observations of the multitude of ingenious methods that people use to satisfy their information needs. Many needs were postponed or not addressed because of attentional cost orcontextual factors. The lack of mobile internet was not the only inibitor. The authors conclude that the device’s sensitivity to the task at hand, situational context, and the links between personal and public data holds promise to ease mobile information access.