NOUVELLES TECHNOLOGIES DE L'INFORMATION ET DE LA COMMUNICATION, NOUVELLES LECTURES, NOUVEAUX ÉCRITS - POUR UNE CARACTÉRISATION DU GENRE DU DOCUMENT NUMÉRIQUE

摘要

Current technologies have enabled knowledge to along pathways that go beyond pluralistic patterns of traditional broadcasting. To improve the indexing and searching in this area, many methods are aimed at overdetermine papers published as a set of metadata. The gendre issue is part of the more global problem of classification of semiotic products and aims to provide elements that could improve research methods and techniques of manipulation.This thesis suggest to highlight discriminating features that could identify the types of digital documents. Ultimately, this should lead to improved search tools and indexing documents. It is organized into four parts. But first, a section introduces the concepts used thoughout the development, it shall determine the epistemological environment and methodologies chosen. The first part of the thesis makes a "flattering" of the concept of document and proceeds with his analysis, particularly in terms of scanning and changing practices that it brings. Three levels were selected as relevant: the technological aspect, witch relates to object (shape), its semiotic consistency, dependent on the interpretation of the subjects (the bottom) and the pragmatic target is it to say that revolve around the actions and justify its existence (it works). One defined the object of study, the second part of the thesis makes a "state of arts" of gender. After tracing the evolution of ideas that the concept of gender has developed over time, were identified three trends that seem to illustrate just as many ways to approach the semiotic object. The third part finally dressed the gendre issue, trying to find items relevant to a closer look operation of the dialectic between "being digital document", "subjects interpretants" and "practices" related. The goal here is to identify discriminating features that could identify the types of digital documents. Gender is thus analyzed:- context as a priori, that to say as a world of expectations of a community, in which case it is a "context of reference" subjects;- buildings as a reading course content, where the "context of actions" provides the defining elements in the formation of a meaning;- as a posteriori description of a document analysis, in this case, the "object context".The fourth section provides an exemplification and discussion of these ideas on a corpus of digital documents. The target application is here. This is to examine the possibility of developing tools for automatic recognition of gendres.