Inria / Raweb 2004
Project-Team: Eiffel

Search in Activity Report, year 2004:


Project-Team : eiffel

Section: New Results

Axis 1: Functional and cooperative activities in design

A current research topic concerns identifying and developing new functionalities to assist cooperation in collective design. The extended enterprise has an increasing need for tools to assist cooperation between people at remote locations in both synchronous and asynchronous modes and to assist the exchange and sharing of information between the various actors involved in the design process. Currently available tools have their limitations and although tools for assisting workflow do exist, there are very few that are dedicated to assisting cooperation. Therefore, research objectives are to define distributed work environments that integrate design tools (e.g. CAD) and communication tools that make use of the possibilities offered by Internet and other means of communication with people on remote locations. To tackle this research topic, we have developed a typology of collective design situations, by identifying the dimensions that characterize these situations. This strong theoretical framework serves to guide research and capitalize on our knowledge of the various collaborative aspects of design. Our research strategy is as follows:

These two types of studies are important at various stages of research and we frequently use these two approaches. The development of cooperative tools is carried out within a framework of collaboration with computer science researchers, both from university and industry.

MAGIE: A web-based cooperative system for product innovation in the automotive industry

Participants: Françoise Darses, Thierry Février Quesada.

The MAGIE project addresses the design issues of a cooperative environment, called COOPARENA. Our design approach is based on cognitive engineering principles, which consist in specifying the cooperative system requirements at a cognitive level but in terms that can be manipulated by the computer engineers in charge of developing the future system. From this standpoint, we first have carried out a cognitive ergonomics analysis of the current collaborative situations, synchronous or asynchronous, in which the team members are involved. This human factor analysis resulted in COOPARENA. This cooperation space is made up of collective tasks (which the partners in the innovation project must necessarily undertake to achieve the process) that are carried out using a finite number of Basic Cooperative Functions. Our approach then consisted in translating this model into UML use cases, in close interaction with the system designers. The resulting architecture of the cooperation environment is thereby able to meet the real needs of the Web-based collaborative platform's future users.

In 2004, the last stage of the MAGIE project gave expression to an assessment deliverable of the implemented system [52]. Two main assessments have been carried out, without and with users. The initial iterative and expert assessment led to the second scenario-based usability test. In a realistic test plan, five stakeholders took part in a 2-days interface evaluation. The task observation and user feedback provided performance measurement. The results show three main gaps in the system. The first one is linked to technical problems, the second to unimplemented functionalities and the last to usability problems. Some positive aspects are also noteworthy. A first one is the ecological advantage of this scenario-based methodology. Second is the encouraging character of this asynchronous cooperative groupware's evaluation. Finally, the training and use of this application appear quick and satisfying in an Internet environment.

On the basis of the MAGIE project (October 2001 - November 2003), Thierry Février Quesada is going to defend a doctoral dissertation (planned for spring 2005). His topic is the collective design activity in innovation projects.

MEDIANNOTE: Annotations in engineering design

Participants: Françoise Darses, Françoise Détienne, Sylvie Guibert.

The aim of this study is to characterize the annotations in product design and to define their role in the process of cooperation between various designers. Two studies have been performed to analyse the practice of annotation in design.

The first study has focused on annotation in co-located design meetings. The method chosen for this study has consisted in observing an experimental design engineering situation: four meetings have been video-recorded. The results show that the annotations made during these meetings are both graphic (and not only textual) and figurative (representations of product components). Thus, they are not only deictic. Moreover, the number of annotations varies according to designers' social function and, more precisely, according to the importance of the product area. Lastly, shared work-space is important for cooperative work. Annotations allow members of a group to focus their attention. Situated in the context of the document, they permit a better understanding of it and provide the other designers with a means of reacting via enrichment or evaluation of the solution, and provision of additional knowledge [53].

The second study focuses on annotation in asynchronous design interactions. In this study, quotation (or citation) in on-line discussions of open-source projects is analyzed as an annotation practice. Content analysis of messages composing design-oriented discussions allows us to better understand the semantics of the links between a quotation and the message-part commenting on it.

A socio-cognitive approach of cooperative design in Open-Source Software (OSS)

Participants: Flore Barcellini, Jean-Marie Burkhardt, Françoise Détienne.

Our research question concerns the design processes of an Open-Source Software (OSS) project devoted to the development of a programming language called Python. The designers of Python engage in a particular process called Python Enhancement Proposals (PEPs). PEPs are akin to a design process, called Request For Comments, that has been practised for decades to define standards for the Internet (used, especially, by the Internet Engineering Task Force). PEPs may also be compared to technical review meetings as practised in many corporate and governmental settings. Thus, our refined research question is as follows: What are the structure and dynamics of PEPs and how do they differ from classical technical review meetings in traditional software development? We have identified two main directions for analysis. Firstly, we are interested in the interaction dynamics of the software designers and implementers. Secondly, we want to understand what we will term the ``socio-technical couplings'' of the OSS project, i.e., the ``statics'' that accompanies the interaction dynamics. The set of methodologies that we have been using are: Social network analysis methods; Corpus-based, computational linguistics and computational measures of stylistics; Discourse analysis and speech act analysis; Ethnography [50]. This work is performed in collaboration with UC Berkeley.

Software reuse and service-based design

Participants: Jean-Marie Burkhardt, Françoise Détienne, Willemien Visser.

Previous empirical studies on design have emphasised the role of memory of past solutions. Design indeed involves the use of both generic knowledge and episodic knowledge about past designs for analogous problems. It thus involves the reuse of past designs. We analyse this reuse from a socio-cognitive viewpoint. According to a purely cognitive approach, reuse involves cognitive mechanisms linked to the problem solving activity. Our socio-cognitive approach accounts for both these phenomena and for reuse linked to cooperation, in particular coordination, and confrontation and integration of viewpoints.

This year, collaboration with the Software Engineering team of Keele University, funded by the French-British fund Alliance, aims to construct and to validate an operational framework for a Software-as-a-Service (SaaS) system, grounded, on the one hand, in the cognitive ergonomics of design activities, in particular reuse in Software Design, and in other relevant user-centred methods and approaches, and, on the other hand, in the methods and concepts of Software Engineering (architecture, function, service, etc.).

Dynamic aspects of design cognition: Elements for a cognitive model of design

Participant: Willemien Visser.

We have formulated a critical review of cognitive design studies, focusing on activities actually implemented in professional, industrial design projects. This analysis leads to elements for a cognitive model of design that, on the one hand, furthers our understanding of design, and on the other hand, offers a basis for the advancement of professional design education and practice. This research [55] is especially concerned with dynamic aspects of design, that is, it focuses on the activity implemented by designers, especially the cognitive processes and/or strategies they use, rather than with static aspects. First we present the two main competing models of today, that is, the symbolic information-processing approach, represented by Herbert A. Simon and the ``situativity'' approach, mainly represented by Donald Schön. Next, we present nuances and critiques with respect to both approaches, and complete and integrate them into our own cognitively oriented dynamic approach to design in which, from a cognitive viewpoint, design is considered to be most appropriately characterized as construction of representations.

Viewpoints in design

Participants: Jean-Marie Burkhardt, Françoise Détienne.

The notion of ``viewpoint'' associated with design activities is here defined as the effects of designers' speciality and possibly their role in the design process on the main constraints and objects that are cognitively favoured. We distinguish the notion of ``viewpoint'' from that of ``representation'', classically used in Cognitive Ergonomics. In order to provide design tools, we present an approach to explore the viewpoints of experts that is based on cognitive-discursive analysis and geometric data analysis [38]. Furthermore, a distinction between three types of viewpoints has been made [33]: prescribed viewpoint, discipline-specific viewpoint and integrated viewpoint. This distinction allows us to analyse the dynamics of viewpoint confrontation and the cooperative modes that enable these different viewpoints to be integrated.