La phase conceptuelle, basée sur des informations imprécises et incomplètes, est à la base de la satisfaction des clients par le futur système. La programmation architecturale se concentre sur une des premières étapes de cette phase dans le domaine de la construction : la définition des besoins d'un bâtiment. Les pratiques et recherches actuelles en Architecture-Ingénierie-Construction ont tendance à être basées sur l'expérience des programmistes et architectes. Elles sont rapidement orientées vers la production de solutions architecturales et avec une formalisation tardive des besoins métiers en exigences bâtiment. La définition des exigences est bien plus développée dans les autres domaines d'ingénierie (i.e. génie mécanique, industriel et logiciel) au sein de théories, modèles, techniques et outils. Ces travaux de recherche visent à proposer un raisonnement de programmation architecturale basé sur ces connaissances en matière d'ingénierie des exigences dans ces autres domaines. La méthodologie suivie pour développer ce raisonnement s'appuie sur une démarche transdisciplinaire basée sur une combinaison de méthodologies issues des Sciences de la Conception. Outre l'identification de langages de modélisation couvrant différent points de vues associées à la programmation architecturale, un manque est identifié en ce qui concerne les concepts définissant l'objet d'étude : la transition entre les concepts d'activité et d'espace. Le concept de ‘‘méta-espace'' est proposé pour combler ce manque ainsi qu'un langage de modélisation permettant l'exploitation des relations entre ces concept. Une formalisation très en amont de la transition des besoins métiers des clients vers les exigences bâtiment est proposée via une démarche conceptuelle résultant de l'ajout du concept de ‘‘méta-espace''. Cette démarche est présentée étape par étape en s'appuyant sur les divers cas d'études traités au cours de la thèse. / The conceptual phase is based on very fuzzy and incomplete information about clients' needs. It constitutes the core of clients' satisfaction about the future system. Architectural programming, or briefing process, focuses on the early stages of the conceptual phase of construction projects, i.e. the requirements definition. Current practices and research in Architecture-Engineering-Construction tend to be experience-based, solution-oriented, and with a late formalisation of business needs into building requirements. The requirements definition is far more supported by theories, models, tools, and techniques in other engineering disciplines (i.e. Mechanical, Industrial, and Software Engineering). This research aims to propose an architectural programming reasoning based on existing requirements engineering knowledge from these other engineering disciplines. The methodology followed to develop this reasoning is a transdisciplinary approach based on a combination of Design Research Methodology, and Design Science. Besides the identification of modelling languages covering different viewpoints associated with the briefing process, a gap is identified regarding modelling constructs associated with a “meta-space” design artefact. This design artefact and its associated modelling language (i.e. the meta-space diagram) support the formalisation of the transition from activities to space. The resulting conceptual framework proposes a step by step early formalisation of the transition from clients' business needs to building requirements. All along the thesis, different case studies are used to illustrate the proposals.

This study aims to design and implement a domain-specific computer programming language: the Building Environment Rule and Analysis (BERA) Language. As a result of the growing area of Building Information Modeling (BIM), there has been a need to develop highly customized domain-specific programming languages for handling issues in building models in the architecture, engineering and construction (AEC) industry. The BERA Language attempts to deal with building information models in an intuitive way in order to define and analyze rules in design stages. The application of the BERA Language aims to provide efficiency in defining, analyzing and checking rules. Specific example applications implemented in this dissertation are on the evaluation of two key aspects: building circulation and spatial programming. The objective of this study is to accomplish an effectiveness and ease of use without precise knowledge of general-purpose languages that are conventionally used in BIM software development. To achieve the goal, this study proposes an abstraction of the universe of discourse - it is the BERA Object Model (BOM). It is a human-centered abstraction of complex state of building models rather than the computation-oriented abstraction. By using BOM, users can enjoy the ease of use and portability of BIM data, rather than complex and platform-dependent data structures. This study also has reviewed and demonstrated its potential for extensibility of BOM. Not only its lateral extensions such as structural building elements, but also the vertical extensions such as additional properties for existing BOM objects are good examples. In current BERA Language Tool, many computed and derived properties/relations have been proposed and implemented, as well as some basic data directly from the given building model. Target users of the BERA Language are domain experts such as architects, designers, reviewers, owners, managers, students, etc., rather than BIM software developers. It means that the people who are interested in the building environment rule and analysis are the potential users. The BERA Language Tool comprises many libraries to alleviate common but unnecessary problems and limitations that are encountered when users attempt to analyze and evaluate building models using commercially available tools. Combined with other libraries which populate rich and domain-specific datasets for certain purposes, the BERA Language will be fairly versatile to define rules and analyze various building environmental conditions.