2Agronomists and computer scientists collaborated to develop a case-based reasoning system in the management of agricultural territories. The case base collects the descriptions of farms from various French regions (Causses, Lorraine). The reasoning consists in re‑using the information contained in the case base to help the analysis of other farms, in connection with specific environmental problems. In the course of this work, agronomists and computer scientists had to answer to various questions that came up out of the modelling process: how should spatial organizations be described and represented? How should the spatial relations between the various areas of a farm be named? How should a common vocabulary for the various types of farms considered be stabilized? How should the resemblance between spatial organizations be measured? How should the knowledge acquired from a farm be adapted and re‑used in the analysis of another farm? The resolution of these questions was supported by a continuous confrontation between the models and reasoning methods used by the agronomists and the computer scientists.

3Psychologists and linguists contributed to help answer these modelling questions. Their approach consisted in the instrumentation and the analysis of the modelling activity performed by the agronomists and computer scientists, based on the linguistic, gestural, and artefactual traces. For that, three work meetings have been video‑taped at three different periods of the project. The linguists and the psychologists, in close connection with the agronomists and the computer scientists, have analyzed the resulting corpus. The whole project lasted from 2001 (first recording) to 2007 (writing of this dossier for the journal).

4The aim of this dossier, entitled “From the sheep house to the computer centre – Collective elaboration of spatial knowledge”, is to relate our interdisciplinary experience in a inter‑related manner. It is composed of five articles, which respond one to each other: the first one is a presentation of the whole project and of its theoretical and practical frameworks. The second and third articles bring together the research issues, the tools for the spatial representation and the models of reasoning, in agronomy and computer science. The fourth one tackles the methodological aspects of the exploitation of the video corpus. A last article attempts to federate all points of view by showing the mutual interest of the partners within the analysis of the video corpus.

5Florence Le Ber, and Christian Brassac, in the first article entitled “A longitudinal study of a knowledge modelling process about farm spatial organizations” introduce the issue in its entirety. They present the framework and the main steps of the ROSA project, which relies on a joint knowledge building process involving both agronomists and computer scientists. The article describes the steps of the modelling process as well as the filming procedures set up by the psychologists and linguists in order to make the underlying knowledge building process explicit and to analyze it.

6In the article entitled “Case‑Based Reasoning and Landscape Agronomy”, Pierre‑Louis Osty, Florence Le Ber, and Jean Lieber, bring together the work of the agronomists, dealing with the organization of farming territories, and the computer modelling method. They show, through the example of the ROSA project, that the meeting of these two research fields can be profitable, for both agronomy, in its search for a qualitative method that better appreciates complexity, and computer sciences, in their approach of new modelling problems.

7Sylvie Lardon, and Mathieu Capitaine, in “Processing and transforming spatial representations for Agronomy” clarify the approach of spatial modelling implemented within the ROSA project. They show how the agronomists engaged in this project have improved their procedures for modelling agricultural spatial organizations and how they have stabilized their modelling tools, and especially the chorems, borrowed from geography. This progression relies on the confrontation between the chorems and the graphs, which are the tools of computer scientists. The article ends up with a methodological proposal for the production and the transformation of spatial representations in agronomy, based on the joint use of these two instruments.

8In the article entitled “Production of Knowledge and Multimodal Interactions”, Lorenza Mondada deals with the video documentation of the joint activities of the agronomists and computer scientists within the ROSA project. She raises two kinds of issues: on one hand, the analysis of scientific and expert practices, including their “praxeological”, situated, embodied and material features, by showing how collective reasoning is intimately embedded both within talk‑in‑interaction and within complex multimodal activities. On the other hand, the emergence of methodological issues which are left rather under-analyzed until now, that concern the analytical use of video data.

9The last article of this dossier, entitled “An Analysis of Knowledge Emergence during a collective process”, is a five‑voices article (Christian Brassac, Sylvie Lardon, Florence Le Ber, Lorenza Mondada, Pierre‑Louis Osty); it reports the concrete interdisciplinary practice which lies in the heart of the ROSA project. Two aspects are investigated: on the one hand, how the collaborative work mobilizes the issues interesting both agronomists and computer scientists; on the other hand, how the twofold presence of human and social sciences (in this case, linguistics and psychology) contributes to a reflexive practice of the actors in their modelling activities. This investigation relies on cross-analysis of a corpus extract, which was selected in common.

10The dossier proposes an interdisciplinary reflection on a particular project, whose members collectively elaborated knowledge on agricultural space. This elaboration process, was performed at several levels, by agronomists and computer scientists, working on knowledge modelling, but also by psychologists and linguists, analysing a posteriori the video corpus. This concrete and interdisciplinary practice of knowledge building is what we want to show to the readers of the Revue d’Anthropologie des Connaissances.