Graphical user interfaces have not changed much since the creation of the Desktop metaphor twenty years ago. Compared with the growth of the computing power of desktop computers, desktop interfaces have been remarkably stable. Interactive applications do not take advantage of new interaction paradigms or interaction techniques. All too often, novel interaction techniques are developed as prototypes and tested individually, while their adoption in real application requires studying their integration with other techniques. We are exploring new programming environment and systems able to quickly integrate new techniques and interaction paradigms as they are discovered. Our activity focuses on creating a new software architecture that will enable new interactive applications to profit from richer graphics and novel interaction paradigms without changing the non-interactive part of the application. We are coordinating the RNTL Indigo project with ILOG, the W3C and the CENA in Toulouse to create new mechanisms based on technologies standardized by the W3C for the infrastructure and from our experience in toolkits for the separation of application and their interaction and presentation. We have also just started to work on an "Interaction Museum", funded by the European Network of Excellence Convivio, to gather existing as well as future interaction techniques so as to make them more readily available to practitioners, researchers and students. This will complement our work on software tools and architectures by exposing the developers of interactive applications to novel techniques, therefore easing the transfer of research results to industry.

Graphical user interfaces (GUIs) are based on a small set of interaction techniques, which rely heavily on two elementary actions: pointing a target on the screen, e.g. an icon or button, and navigating to a non-visible part of the information space, e.g. by scrolling or zooming.

We are working on improving pointing and navigation performance in GUIs. Indeed, the performance of pointing on a screen is similar to that of pointing in the physical world, and it should be possible to take advantage of the computing power to get a significant advantage when pointing in an information world. The major theoretical tool to study pointing performance is Fitts' law , which defines the movement time as an affine function of the index of difficulty (ID), defined as the log of the ratio between target distance and target width. In other words, pointing performance strictly depends on the relative size of the target to the distance to the target.

We have explored a technique called target expansion which grows the size of the target when the cursor is near it, and showed that the index of the difficulty of the task is that of the expanded target, even when expansion occurs when the cursor has traveled 90% of the distance to the target. However taking advantage of this property proves to be difficult because it requires a proper anticipation by the system of the target to be expanded. Indeed, expanding the wrong target impairs performance and goes against our primary goal.

We have developed two techniques that look for more promising. The first one is semantic pointing. Semantic pointing uses two independent sizes for each potential target presented to the user: one size in motor space adapted to its importance for the manipulation, and one size in visual space adapted to the amount of information it conveys. This decoupling between visual and motor size is achieved by changing the control-to-display ratio according to cursor distance to nearby targets.

A target with a small visual size and a large motor size will display little information but will be easy to select, which makes it appropriate for, e.g., buttons or links in a Web Page. A target with a large visual size and a small motor size will display more information but will be hard to select, which is appropriate for non-interactive informative labels, for example. A controlled experiment has confirmed that the index of difficulty that best predicts pointing time is the one that uses the motor size rather than the visual size. A prototype application shows how this technique applies to standard GUIs, e.g. scrollbars, dialog boxes, menus and web pages.

The second technique is called vector pointing. It can be seen as an extreme version of semantic pointing where the cursor jumps ``empty'' space and moves from target to target with very little mouse motion. We have shown with a control experiment that Fitts' law does not apply. Indeed, the pointing time of a single target with semantic pointing is constant! In a real application, the performance of vector pointing depends on the accuracy of the prediction of the target being aimed by the user and the influence of distractors. Work is ongoing to develop an application for testing vector pointing in a real setting.

We have also persued our work on multiscale navigation, i.e. navigation of an information world that can be zoomed in or out at any scale (also called Zoomable User Interface or ZUI). Following up on our work that showed that Fitts' law applies to tasks with an extremely high level of difficulty (30 or more, i.e. a ratio of 2³⁰ between target size and target distance) we have developed a theoretical model to explain this result. The model shows an effect of view size on multiscale navigation, which we successfully tested in a controlled experiment. We will continue this work in the context of the Micromegas project (see section ) where we will develop ZUIs and study novel navigation techniques.

The complexity of an interaction technique for a given task, i.e. a given interaction sequence, measures the cost of the actions relative to the size of the task when using this interaction technique. The work on pointing and navigation studies the limit performance of human subjects in such tasks. However it is often difficult to observe such performance when the technique is used in the context of a real application. In order to better understand how interaction techniques behave in context, we are developing a model and associated tool to describe interaction techniques and predict their comparative performance for multiple tasks representative of different interaction contexts. The model, called CIS (Complexity of Interaction Sequences) introduces the notion of complexity for interaction techniques. We have successfully tested the model by evaluating 3 interaction techniques (fixed palettes, bimanual palettes and toolglasses) and shown that the most efficient technique depends on the interaction context, confirming the results of one of our earlier studies . We intend to develop CIS in several directions like improving predictions precision and automating the identification of best and worst contexts for an interaction technique.

Most computer software for managing on-line documents assumes that all user interaction takes place via a keyboard, mouse and screen and that documents exist only in their on-line form. Yet users often write by hand and annotate printed documents, forcing them to juggle on-line and off-line versions. We are exploring the concept of ``interactive paper'' , with the goal of better integrating physical and computer documents and allow users to take advantage of the best aspects of each.

We have been working for several years with research biologists at the Pasteur Institute, observing their use of laboratory notebooks. For both practical and legal reasons, lab notebooks must exist in a physical format, even though a large percentage of the information they contain may come from a computer. From a user's perspective, some information is best entered by hand as ink on paper, e.g. making annotations or quick sketches, whereas other information is best entered on a computer, e.g. generating a graph of the results of an experiment. We have collaboratively designed several prototypes that help biologists do both, maintaining a physical notebook while simultaneously generating an interactive, on-line version.

This year, we began working with a new technology, the Anoto pen, which allows us to capture the strokes as the user writes on ordinary paper. This provides an extremely light-weight method of obtaining a copy of the hand-written parts of the notebook, which we can then provide a variety of on-line services. We developed a working prototype in collaboration with several biologists and plan to test it with them in 2004. Of particular interest is investigating how best to let the biologists create their own personal annotation systems, which they can then use to create personalized indexes, facilitate searching, and provide links to both physical and on-line objects.

In the past two months, we began two new projects in a related area, which involves working with historical manuscripts. For historians, the physical characteristics of these manuscripts are often as important as the information they contain. They need to interact with the actual physical object (often without touching it) while simultaneously entering information into a computer, for subsequent analysis. We are interested in exploring new techniques for linking physical and on-line documents, as well as providing novel methods of visualizing their contents. We are also working with archivists, who are interested in identifying and visualizing the evolution of related documents over time. We ran an initial workshop with historians and archivists in November, and will continue collaborating closely with them next year.

Our research on Information Visualization is organized around three main topics: infrastructure, evaluation and applications.

Over the recent years, technology has played an increasing role in distant communication between people. The growing number of computer systems, combined with the increase of network speed allows us to exchange a variety of data such as text, sound, images or video, in both a synchronous or asynchronous way. However, many of the tools and protocols we use were initially designed for one-to-one communication and quickly show their limits when used for group communication. As an example, gathering pictures of a wedding from various sources via email, ftp or http and making them available on the Internet can be really cumbersome.

In the context of this work, we are mostly interested in group-based communication between close-knit groups of people. The close-knit relationships to which we refer are close to the concept of intimate social network. Knowing someone, working with that person or being part of the same family does not necessarily imply that kind of intimate relationship. Although intimate social networks are omnipresent in our daily life, they are very poorly supported by current communication systems.

We are currently developing Circa, a software infrastructure to support communication within intimate social networks. Circa was designed with two goals in mind: offering users a way to share and exchange data with small groups that they can easily create and manage and providing developers with tools allowing their applications to benefit from this infrastructure.

We are also exploring ways to combine Circa with other software infrastructures developed by colleagues from PUC-Rio University (Rio de Janeiro, Brazil). Common interests include support for user mobility as well as context-aware services, from both a user and a system perspective (i.e. services that depend on the user activity or on the computing resources available).

One of the advantages of video over audio for mediated communication is the ability to transmit non-verbal information. Physical proximity between people is a language for non-verbal communication that we all employ everyday, although we are barely aware of it. Yet, existing systems for video-mediated communication fail to fully take into account these proxemics aspects of communication.

The MirrorSpace project aims at creating an original personal video communication system that takes physical proximity into account. Whereas existing systems aim at creating a single shared space corresponding to a particular interpersonal distance, the goal of MirrorSpace is instead to create a continuum of space that will allow a variety of interpersonal relationships to be expressed. Our work focuses on the understanding of how people's interactions can trigger smooth transitions between situations as extreme as general awareness of remote activity where anonymity is preserved to intimate situations where people can look into the eyes of a remote person.

MirrorSpace units (see Fig. ) combine a digital camera, a flat screen and a proximity sensor. As we aim to support intimate forms of communication, it felt important to us that people could actually look into each other's eyes and possibly merge their portraits into one, so the camera was placed right in the middle of the screen. This setup allows participants to come very close to the camera while still being able to see the remote people and interact with them.

The proximity sensor that measures the distance to the closest object or person in front of it. This distance is used by MirrorSpace software to alter the remote images displayed, and possibly the local one. A blur filter is applied on the images to visually express a distance computed from the local and remote sensor values. Blurring distant objects and people in MirrorSpace allows one to perceive their movement or passing with a minimum involvement. It also offers a simple way of initiating or avoiding a change to a more engaged form of communication by simply moving closer or further away.

This work started as part of the interLiving project (IST/FET, Disappearing Computer initiative, 2000-2003). MirrorSpace was first exhibited in a public setting in February 2003, at Jeune Création, a contemporary art exhibition in Paris. It was then exhibited at Mains d'Oeuvres (Saint-Ouen) (May 2003), Pas vu, pas pris (July 2003) and at the Interactive Design exhibit in the Pompidou Center (Paris December 2003 - February 2004). It was also featured in two national newspapers : Libération (20 February, 2003) and Les Echos (15 December, 2004). See http://insitu.lri.fr/~roussel/projects/mirrorSpace/ for more details.