Literatura científica selecionada sobre o tema "Programmation certifiée"

Objective: To describe a pedagogical method using an Audience Response System (ARS), commonly referred to as “clickers,” for use in preparing athletic training students (ATSs) for the Board of Certification (BOC) exam. Background: The BOC is the only certifying agency for athletic trainers in the United States. ATSs may become a certified athletic trainer (AT) upon graduating from a Commission on Accreditation of Athletic Training Education (CAATE) accredited Athletic Training Education Program (ATEP) and successfully passing the BOC exam. In 2012, CAATE standards were revised to require programs to publish their graduates' recent exam scores on their website (in aggregate form). Description: Preparation for the BOC exam can be a monumental and stress-inducing task for the ATS due to the plethora of information to review. Most ATEPs include test preparation assistance within the curriculum. To date, a review of the literature revealed no published documentation of successful test preparation strategies. Clinical Advantages: The data collected from use of the ARS during BOC exam preparation can be a valuable tool in identifying domain areas in which an ATS may be deficient. In doing so, an ATS may redirect study time to focus on weaker concepts. Additionally this data, when trended, can provide invaluable programmatic information when used as part of a larger assessment plan. As part of the revised CAATE standards, programs are also required to meet or exceed a score of 70% for the first-time pass rate. As a result, there may be a heightened need for test preparation strategies. Conclusion: Athletic training educators seeking a pedagogical method for BOC exam preparation may benefit from the implementation of an ARS to increase ATS motivation and accountability. As an additional benefit, programmatic educational assessment data may also be collected.

Four significant developments have occurred that amplify opportunity for certified organic growers to grow highbush blueberry (Vaccinium corymbosum) successfully. First, there is the 2002 U.S. Department of Agriculture national organic standard that defines organic production practices and crop labels that creates clarity and evens competition. Second, we have the continued increase of smallfruit and vegetable sales related to nutritional and human health reasons and the related market perception valuing organic produce more highly. Third, new tools are becoming available to organic growers that reduce the risk from pest problems such as the recent Organic Materials Review Institute listing of spinosad registered as a wettable powder (Entrust) and a fruit fly bait (GF-120 NF Naturalyte). Finally, the Rutgers Blueberry Working group has made considerable progress in refining integrated pest management practices and in developing new tools for organic production systems. This “work-in-progress” is investigating alternative approaches to some current agricultural practices in soil building, fertility, cultural approaches, and pest management. The authors' 7-year program has demonstrated organic methods in managing new sources of mulch, two key insect pests, two common diseases, and several weed species in establishing a commercial organic production system for highbush blueberries. As a programmatic result, organic acreage in New Jersey has increased from 0 to more than 150 acres, and more than 40 organic growers have adopted parts of this holistic production system in North America.

39 Background: The mission of the Prostate Cancer Supportive Care (PCSC) Program at the Vancouver Prostate Centre (VPC) is to provide clinical care focused on the needs of prostate cancer patients and partners. It is comprised of six modules that are administered by medical professionals in urology, radiation oncology, sexual health nurses, registered dieticians, certified exercise physiologist, male pelvic floor physiotherapists, and couples’ counselling. Incorporation of research into daily care provides evidence for these practices, identifies areas for improvement, and tests new approaches. In order to evaluate the interface between the clinical and research programs, we reviewed the metrics of our PCSC program. Methods: Research studies were grouped by type. Screening and enrollment logs were reviewed to tally the total number of patients approached versus enrolled. Reasons for non-participation based on data in our enrollment logs were categorized. Results: Between Feb 2015 and Mar 2019, PCSC Program participated in 22 research studies: 9 therapeutic or lifestyle intervention studies (3 RCTs), 3 observational studies, 2 registries, 1 survey, 1 genetic study, 1 databank , 2 collaborative programmatic studies, 3 “permission to contact” studies (referral to the study team only). 8 of the 22 studies included recruitment of dyads (both patient and their partner or caregiver). Of 1080 consenting patients, 760 (70.4%) enrolled in 1 study, 210 (19.4%) in 2 studies, and 110 (10.2%) in > 3 studies. 583 patients did not consent due to lack of interest (43.7%), not available (21.1%), time constraints (10.3%), travel distance (6.7%). Conclusions: Our data show that a subspecialty supportive care program can provide a rich environment in which to conduct clinical research. We believe that the integration of the research program and personnel into the clinical setting is key to our success. Current on-going studies are evaluating the impact of and patient satisfaction with all PCSC modules.

Context The clinical immersion experience is a key part of the socialization of the professional athletic training student. Clinical immersion offers the student the chance to experience the totality of the role of the athletic trainer. Programmatic autonomy allows many athletic training educators the opportunity to be creative in the implementation of clinical immersion(s). Objective Examine clinical immersion from the preceptors’ perspective to understand their opinions regarding the structure of the experience and preferences they may have surrounding clinical immersion structure. Design Qualitative exploratory case study design. Setting Individual one-on-one video-conference interview. Patients or Other Participants Eight athletic training preceptors (6 female, 2 male) completed one-on-one interviews and were 38 ± 6 years old with 10 ± 3 years of experience as a preceptor and 14 ± 7 years of experience as a certified athletic trainer. Data Collection and Analysis Semistructured interviews were conducted with all preceptors. Interviews were recorded and transcribed using a video-conferencing software. An iterative approach with principles of phenomenological research was used to code the data. Data saturation guided recruitment and cessation of new interviews. Basic member checks, reflexivity, and multiple analyst triangulation were used to determine trustworthiness. Results The following four major themes emerged: (1) preceptors prefer the clinical immersion experience to be longer than the minimum accreditation requirement, (2) preceptors prefer to supervise second-year students in the immersion experience, (3) immersion allows for meaningful relationship development between the preceptor and student, and (4) preceptors treat immersions as a transition to practice mechanism. Conclusions Preceptors perceived the clinical immersion experience to be a facilitator of strong preceptor-student relationships and a facilitator of transition to practice. They believed that the clinical immersion should be longer than 4 weeks and occur late in the athletic training curriculum.

Background Men experience twice the mortality of women while on ART in sub-Saharan Africa (SSA) largely due to late HIV diagnosis and poor retention. Here we propose to conduct an individually randomized control trial (RCT) to investigate the impact of three-month home-based ART (hbART) on viral suppression among men who were not engaged in care. Methods and design A programmatic, individually randomized non-blinded, non-inferiority-controlled trial design (ClinicalTrials.org NCT04858243). Through medical chart reviews we will identify “non-engaged” men living with HIV, ≥15years of age who are not currently engaged in ART care, including (1) men who have tested HIV-positive and have not initiated ART within 7 days; (2) men who have initiated ART but are at risk of immediate default; and (3) men who have defaulted from ART. With 1:1 computer block randomization to either hbART or facility-based ART (fbART) arms, we will recruit men from 10–15 high-burden health facilities in central and southern Malawi. The hbART intervention will consist of 3 home-visits in a 3-month period by a certified male study nurse ART provider. In the fbART arm, male participants will be offered counselling at male participant’s home, or a nearby location that is preferred by participants, followed with an escort to the local health facility and facility navigation. The primary outcome is the proportion of men who are virally suppressed at 6-months after ART initiation. Assuming primary outcome achievement of 24.0% and 33.6% in the two arms, 350 men per arm will provide 80% power to detect the stated difference. Discussion Identifying effective ART strategies that are convenient and accessible for men in SSA is a priority in the HIV world. Men may not (re-)engage in facility-based care due to a myriad of barriers. Two previous trials investigated the impact of hbART on viral suppression in the general population whereas this trial focuses on men. Additionally, this trial involves a longer duration of hbART i.e., three months compared to two weeks allowing men more time to overcome the initial psychological denial of taking ART.

Background: Certification of hospitals as advanced primary stroke centers (PSC) and comprehensive stroke centers (CSC) leads to improvements in stroke care. We aimed to characterize the specialization and demographics of programmatic leadership at stroke centers nationwide, and to determine the proportion of programs lead by Vascular Neurologists. Methods: We identified hospitals holding advanced primary stroke center and comprehensive stroke center designation using the Joint Commission, DNV, and HFAP websites as of September 2013. The identity of the stroke center director was obtained by contacting each individual hospital via their website or telephone. Specialization of each director was confirmed using data available from relevant medical subspecialty board. Gender and date of graduation from medical school was obtained from an online physician database (Healthgrades). Results: There were 1167 PSC and 50 CSC hospitals in the United States certified by the JC (N=1121), DNV (67) and HFAP (35). The director, confirmed at 940 of 1217 (77%) centers, was most often a neurologist (N=747, 81%) followed by Emergency Medicine (58), internal medicine (17), radiology (15), neurosurgery (13) and psychiatry (11). Neurologists with added certification in vascular neurology (n=319) made up just over one third of identified certified stroke center directors. Certified stroke center directors were most often male (N=764, 81%) and experienced, with a median year of graduation from medical school of 1990 (IQR 1981-1997). Among vascular neurologists 84% were male and the median year of medical school graduation was 1991 compared to 1988 for non-vascular neurologists (p=0.023 by Mann-Whitney). A higher proportion of Medical Directors in CSCs were board certified in Vascular Neurology (74%). Conclusions: Most certified stroke centers were lead by a neurologist, and a third by Vascular Neurologists. Certified stroke center directors are predominantly male, with over 20 years of post-medical school experience.

While movement toward the education of students with disabilities in integrated physical education has now become common internationally, it is not without concerns. Notably, scholars have questioned whether instruction in integrated physical education settings provides inclusive experiences for students with disabilities. The purpose of this exploratory study was to explore certified adapted physical educators’ (CAPEs’) perspectives on barriers and facilitators to students with disabilities experiencing inclusion in integrated physical education. A sample of 99 CAPEs (74.7% female) across the United States provided valid responses to an online survey form. Guided by the established two-step coding protocol, two independent reviewers coded the responses to open-ended questions. In total, 460 barriers and 473 facilitators that fell into seven categories were identified. Among these, both teacher-related barriers and facilitators were more commonly reported than other factors, such as environmental and policy barriers/facilitators. As such, it appears that the participants placed onus on ensuring the inclusiveness of integrated physical education classes largely on the shoulders of the teachers. In addition, codes related to programmatic or equipment-related factors were underreported, suggesting that even in integrated physical education classes where adapted equipment (i.e. equipment facilitator) and personnel support (i.e. programmatic facilitator) are available, inclusive experience may not be. These findings suggest inclusiveness of integrated physical education is complex and influenced by numerous factors.

This article analyzes the strategic potential and empirical challenges of private governance in promoting decent work in global agriculture by curtailing buyer power and fostering labor agency, drawing lessons from Fairtrade International-certified flower plantations in Ecuador. The study explains (1) Fairtrade’s logic in promoting ‘trade fairness’ and ‘worker empowerment’ and operationalization of these values via its certification standards, (2) the power relations shaping certification practices in global flower markets and Ecuadorian plantations, and (3) the grounded implications of participation for firms and workers. Important lessons emerge. First, while Fairtrade pursues a promising avenue for challenging the buyer control that erodes supplier and worker power, it has only marginally reduced floral buyer power due to retailer resistance, low-bar certification competition, and programmatic regulatory gaps. Second, although Fairtrade’s empowerment approach has strengthened labor agency within and beyond the workplace, bolstering individual empowerment has proved easier than fostering associational power. As I show, certification practices and outcomes are mediated by commodity-specific global market politics and localized enterprise, labor force, and legal contestations which explain why program aspirations are often not realized. Standard systems can reshape internal trade relations and organizations but cannot alone ensure global trade equity or robust labor representation.

One of the most important developments in tourism and hospitality education in the last decades has been the implementation of institutional and programmatic international accreditation systems to support and ensure the competitiveness and excellence levels of the education centres and the study paths they offer. Institutions in the ASEAN region can (and should) opt for these international accreditations if they aim to take profit of the various advantages they offer, but how prepared are its centres nowadays to obtain these types of certifications? In order to answer this question, the aim of this paper is to examine and critically analyse the present situation of tourism and hospitality undergraduate programs in Thailand, as an example of the ASEAN context. To carry out this evaluation, a questionnaire was directed to executive staff at the institutions offering undergraduate international programs in Thailand, and to the alumni of those institutions, based on the guidelines of the UNWTO TedQual certification. The findings of this research show that despite the overall positive result (form the perspective of both the alumni and the university staff) there are some aspects which are subject to improvements in order to enhance the performance of the programs and their possibilities of being certified with international accreditations in tourism and hospitality.

Although trachoma mass drug administration (MDA) programs target ocular Chlamydia trachomatis, the global trachoma control program does not monitor infection as a measure of impact but instead relies on monitoring clinical indicators. This study aimed to monitor the prevalence of ocular C. trachomatis among a population-based sample of children ages 1–5 years throughout Amhara, Ethiopia, a region that has received approximately 8 years of annual MDA as part of trachoma control. Between 2014 and 2021, trachoma impact surveys and surveillance surveys were conducted in all 156 districts of Amhara using a multistage cluster randomized methodology. Certified graders assessed individuals ages ≥ 1 year for trachomatous inflammation-follicular (TF), and a random subset of children ages 1–5 years also provided a conjunctival swab. Polymerase chain reaction was used to test for C. trachomatis. A total of 28,410 conjunctival swabs were collected from children ages 1–5 years across Amhara. The regional C. trachomatis infection prevalence was 4.7% (95% uncertainty interval: 4.3–5.1%). Infection was detected in all 10 zones of the region and ranged from 0.2% in Awi Zone to 11.9% in Waghemra Zone. Infection was detected in 17 (26%) districts with a TF prevalence < 10% and in 7 (21%) districts with a TF prevalence < 5%. Through programmatic monitoring of C. trachomatis infection, this study demonstrated that considerable infection remained throughout Amhara despite approximately 8 years of trachoma interventions and that enhanced interventions such as more frequent than annual MDA will be needed if elimination thresholds are to be reached.

Nous nous intéressons ici à la génération de programmes certifiés
corrects par construction. Ces programmes sont obtenus en
extrayant l'information pertinente de preuves constructives réalisées
dans l'assistant de preuves Coq.

Une telle traduction, ou "extraction", des preuves constructives
en programmes fonctionnels n'est pas nouvelle, elle correspond
à un isomorphisme bien connu sous le nom de Curry-Howard. Et
l'assistant Coq comporte depuis longtemps un tel outil d'extraction.
Mais l'outil précédent présentait d'importantes limitations. Certaines
preuves Coq étaient ainsi hors de son champ d'application, alors que
d'autres engendraient des programmes incorrects.

Afin de résoudre ces limitations, nous avons effectué une refonte
complète de l'extraction dans Coq, tant du point de vue de la théorie
que de l'implantation. Au niveau théorique, cette refonte a entraîné
la réalisation de nouvelles preuves de correction de ce mécanisme
d'extraction, preuves à la fois complexes et originales. Concernant
l'implantation, nous nous sommes efforcés d'engendrer du code
extrait efficace et réaliste, pouvant en particulier être intégré dans des
développement logiciels de plus grande échelle, par le biais de
modules et d'interfaces.

Enfin, nous présentons également plusieurs études de cas illustrant
les possibilités de notre nouvelle extraction. Nous décrivons ainsi la
certification d'une bibliothèque modulaire d'ensembles finis, et
l'obtention de programmes d'arithmétique réelle exacte à partir d'une
formalisation d'analyse réelle constructive. Même si des progrès
restent encore à obtenir, surtout dans ce dernier cas, ces exemples
mettent en évidence le chemin déjà parcouru.

Cette thèse s’intéresse à la programmation certifiée correcte dans le cadre formel fourni par l’assistant de preuve Coq, et conduite par étapes de raffinements, avec l'objectif d’aboutir à un résultat correct par construction. Le langage de programmation considéré est un langage impératif simple, avec affectations, alternatives, séquences, et boucles. La sémantique associée à ce langage est une sémantique relationnelle exprimée dans un cadre prédicatif plus adapté à un plongement dans la théorie des types, plutôt que dans le calcul des relations. Nous étudions la relation entre d’une part la sémantique prédicative et relationnelle que nous avons choisie, et d’autre part une approche plus classique dans le style de la logique de Hoare. En particulier, nous montrons que les deux approches ont en théorie la même puissance. La démarche que nous étudions consiste à certifier, avec l’aide d’un assistant de preuve, les raffinements successifs permettant de passer de la spécification au programme. Nous nous intéressons donc aussi aux techniques de preuve permettant en pratique d’établir la validité des raffinements. Plus précisément, nous utilisons un calcul de la plus faible pré-spécification jouant ici le rôle du calcul de la plus faible pré-condition dans les approches classiques. Afin que l’articulation des étapes de raffinement reste aussi proche que possible de l’activité de programmation, nous formalisons un langage de développement qui permet de décrire l’arborescence des étapes de raffinement, ainsi qu’une logique permettant de raisonner sur ces développements, et de garantir leur correction
This thesis investigates certified programming by stepwise refinement in the framework of the Coq proof assistant. This allows the construction of programs that are correct by construction. The programming language that is considered is a simple imperative language with assignment, selection, sequence, and iteration. The semantics of this language is formalized in a relational and predicative setting, and is shown to be equivalent to an axiomatic semantics in the style of a Hoare logic. The stepwise refinement approach to programming requires that refinement steps from the specification to the program be proved correct. For so doing, we use a calculus of weakest pre-specifications which is a generalisation of the calculus of weakest pre-conditions. Finally, to capture the whole refinement history of a program development, we formalize a design language and a logic for reasoning about program designs in order to establish that all refinement steps are indeed correct. The approach developed during this thesis is entirely mecanised using the Coq proof assistant

Les langages synchrones sont apparus autour des années quatre-vingt, en réponse à un besoin d'avoir un modèle mathématique simple pour implémenter des systèmes temps réel critiques. Dans ce modèle, le temps est découpé en instants discrets durant lesquels tous les composants du système reçoivent et produisent une donnée. Cette modélisation permet des raisonnements beaucoup plus simples en évitant de devoir prendre en compte le temps de calcul de chaque opération. Dans le monde du logiciel critique, la fiabilité du matériel et de son fonctionnement sont primordiaux, et on accepte d'être plus lent si on devient plus sûr. Afin d'augmenter cette fiabilité, plutôt que de concevoir manuellement tout le système, on utilise des machines qui synthétisent automatiquement le système souhaité à partir d'une description la plus concise possible. Dans le cas du logiciel, ce mécanisme s'appelle la compilation, et évite des erreurs introduites par l'homme par inadvertance. Elle ne garantit cependant pas la bonne correspondance entre le système produit et la description donnée. Des travaux récents menés par une équipe INRIA dirigée par Xavier Leroy ont abouti en 2008 au compilateur CompCert d'un sous-ensemble large de C vers l'assembleur PowerPC pour lequel il a été prouvé dans l'assistant de preuve Coq que le code assembleur produit correspond bien à la description en C du programme source. Un tel compilateur offre des garanties fortes de bonne correspondance entre le système synthétisé et la description donnée. De plus, avec les compilateurs utilisés pour le temps réel critique, la plupart des optimisations sont désactivées afin d'éviter les erreurs qui y sont liées. Dans CompCert, des optimisations elles aussi prouvées sont proposées, ce qui pourrait permettre ces passes dans la production de systèmes temps réel critiques sans en compromettre la fiabilité. Le but de cette thèse est d'avoir une approche similaire mais spécifique à un langage synchrone, donc plus approprié à la description de systèmes temps réel critiques que ne l'est le C. Un langage synchrone flots de données semblable à Lustre, nommé Ls, et un langage impératif semblable au langage C, nommé Obc y sont proposés ainsi que leur sémantique formelle et une chaîne de compilation avec des preuves de préservation de sémantique le long de cette chaîne.

En 1974, Boyce et Chamberlin ont créé le langage SQL en se basant sur l'algèbre relationnelle proposée par Codd en 1970, mais à mesure d'extensions, la sémantique formelle de SQL s'est éloignée de celle de l'algèbre relationnelle. Le petit fragment select from where de SQL peut correspondre à une algèbre relationnelle avec une sémantique multiensemble en restreignant les expressions et les formules à celles exprimables en algèbre relationnelle. Pour capturer la sémantique du fragment beaucoup plus réaliste select from where group by having en prenant en compte toutes les expressions y compris celles avec agrégats, toutes les formes de formules, les valeurs nulles et, encore plus subtil, les environnements très particuliers de SQL, Benzaken et Contejean ont proposé l'algèbre SQLalg qui est une extension de l'algèbre relationnelle avec un nouvel opérateur pour la partie group by having conçu spécifiquement pour prendre en compte tous les aspects de SQL cités précédemment. Ce même fragment de SQL, avec toute ses subtilités, est-il capturable par l'algèbre relationnelle imbriquée ? Cette thèse prouve formellement que oui. En effet, nous proposons une traduction, certifiée en coq, de SQLalg vers NRAᵉ, qui est une formalisation en coq de l'algèbre relationnelle imbriquée. La traduction prend en compte les expressions simples et complexes, les formules SQL et reflète parfaitement comment les environnements sont construits et manipulés, spécialement pour les agrégats et les requêtes corrélées. Ce travail s'inscrit dans un cadre plus global, celui du projet DBCert: une chaîne de compilation certifiée en Coq de SQL vers JavaScript
In 1974, Boyce and Chamberlin created sql using the concepts of the relational algebra proposed by Codd in 1970, but as it evolved, its formal semantics became more and more complex. The small fragment select from where of SQL can be mapped to a relational algebra, with bag's semantics and by restricting expressions and formulae to those which can be expressed in relational algebra. To capture the semantics of the much more realistic fragment select from where group by having, taking into account all the expressions including those with aggregates, all forms of formulas, null values and, even more subtle, the specific SQL's environments, Benzaken and Contejean propose SQLalg which is an extension of relational algebra with a new operator for the group by having part designed specifically to take into account all the aspects of SQL cited above. Can this same fragment of SQL, with all its subtleties, be captured by nested relational algebra ? The present work formally proves that the answer to this question is yes. Indeed, we have built a certified translation, formalized in Coq, from SQLalg to NRAᵉ, which is a formalization in Coq of the nested relational algebra. The translation supports simple and complex expressions, SQL's formulae and perfectly reflects how environments are built and manipulated, especially for aggregates and for correlated queries. This work is part of a more global framework : the DBCert project which is a compilation chain certified in Coq from SQL to JavaScript

An end-to-end guarantee of software correctness by formal verification must consider two sources of bugs. First, the verification tool must be correct. Second, programs are often verified at the source level, before being compiled. Hence, compilers should also be trustworthy. Verifiers and compilers' complexity is increasing. To simplify code analysis and manipulation, these tools rely on intermediate representations (IR) of programs, that provide structural and semantic properties. This thesis gives a formal, semantic account on IRs, so that they can also be leveraged in the formal proof of such tools. We first study a register-based IR of Java bytecode used in compilers and verifiers. We specify the IR generation by a semantic theorem stating what the transformation preserves, e.g. object initialization or exceptions, but also what it modifies and how, e.g. object allocation. We implement this IR in Sawja, a Java static analysis toolbench. Then, we study the Static Single Assignment (SSA) form, an IR widely used in modern compilers and verifiers. We implement and prove in Coq an SSA middle-end for the CompCert C compiler. For the proof of SSA optimizations, we identify a key semantic property of SSA, allowing for equational reasoning. Finally, we study the semantics of concurrent Java IRs. Due to instruction reorderings performed by the compiler and the hardware, the current definition of the Java Memory Model (JMM) is complex, and unfortunately formally flawed. Targetting x86 architectures, we identify a subset of the JMM that is intuitive and tractable in formal proofs. We characterize the reorderings it allows, and factor out a proof common to the IRs of a compiler.

Les applications informatiques sont de plus en plus présentes dans nos vies. Pour les applications critiques (médecine, transport, . . .), les conséquences d’une erreur informatique ont un coût inacceptable, que ce soit sur le plan humain ou financier. Une des méthodes pour éviter la présence d’erreurs dans les programmes est la vérification déductive. Celle-ci s’applique à des programmes écrits dans des langages de haut-niveau transformés, par des compilateurs, en programmes écrits en langage machine. Les compilateurs doivent être corrects pour ne pas propager d’erreurs au langage machine. Depuis 2005, les processeurs multi-coeurs se sont répandus dans l’ensemble des systèmes informatiques. Ces architectures nécessitent des compilateurs et des preuves de correction adaptées. Notre contribution est l’extension modulaire d’un compilateur vérifié pour un langage parallèle ciblant des architectures parallèles multi-coeurs. Les spécifications des langages (et leurs sémantiques opérationnelles) présents aux divers niveaux du compilateur ainsi que les preuves de la correction du compilateur sont paramétrées par des modules spécifiant des éléments de parallélisme tels qu’un modèle mémoire faible et des notions de synchronisation et d’ordonnancement entre processus légers. Ce travail ouvre la voie à la conception d’un compilateur certifié pour des langages parallèles de haut-niveau tels que les langages à squelettes algorithmiques
Nowadays, we are using an increasing number of computer applications. Errors in critical applications (medicine, transport, . . .) may carry serious health or financial issues. Avoiding errors in programs is a challenge and may be achieved by deductive verification. Deductive verification applies to program written in a high-level languages, which are transformed into machine language by compilers. These compilers must be correct to ensure the nonpropagation of errors to machine code. Since 2005, multicore processors have spread in all electronic devices. So, these architectures need adapted compilers and proofs of correctness. Our work is the modular extension of a verified compiler for parallel languages targeting multicore architectures. Specifications of these languages (and their operational semantics) needed at all levels of the compiler and proofs of correctness of this compiler are parameterized by modules specifying elements of parallelism such as a relaxed memory model and notions of synchronization and scheduling between threads. This work is the first step in the conception of a certified compiler for high-level parallel languages such as algorithmic skeletons

Cette these decrit un systeme pour specifier, programmer et certifier (prouver) des bibliotheques de calcul formel. Chaque unite de bibliotheque represente une certaine structure algebrique, comme les groupes, les anneaux, les algebres de polynomes, etc. En pratique, les bibliotheques de calcul formel comportent un grand nombre de structures algebriques qui partagent entre elles algorithmes, proprietes et preuves, de facon assez compliquee. Par exemple, une notion calculatoire d'algebre (comme le pgcd) est definie par des proprietes algebriques. Toutes les structures algebriques verifiant ces proprietes partagent cette notion. Mais chacune d'elles peut l'implanter en utilisant differents algorithmes : generiques (communs a une large gamme de structures), ou specialises (utilisant une representation particuliere des donnees, ou communs a des structures verifiant une certaine propriete). Pour exprimer ce partage entre unites, l'utilisateur dispose de mecanismes d'heritage. L'heritage permet de creer une unite a partir de une ou plusieurs autres, en decrivant uniquement les differences de l'unite creee vis-a-vis des unites heritees, tout en preservant leurs proprietes. En effet, les algorithmes (et les preuves) applicables a une unite a seront aussi applicables a une unite b qui herite de a. Par rapport aux mecanismes traditionnels de programmation orientee objet, l'heritage est ici tres restreint de maniere a garantir la coherence logique des bibliotheques et la terminaison des programmes. Ce systeme est d'abord implante en coq. Il est ensuite decrit de maniere abstraite (independament de toute syntaxe), dans un metalangage categorique, en s'inspirant des categories contextuelles introduites par cartmell pour modeliser les types dependants. Pour donner lieu a des programmes executable, il est aussi propose une traduction de ces bibliotheques dans le langage ocaml, en utilisant divers traits de ce langage : polymorphisme, objets, modules, etc.

Cette thèse porte sur la construction d'un environnement pour développer des librairies de calcul formel certifié. Nous présentons d'abord les espèces, structures servant à décrire des spécifications par héritage multiple, raffinement et paramétrisation. Les collections, construites par encapsulation d'espèces constituent la librairie utilisateur. Nous définissons également les analyses statiques garantissant la correction d'une définition d'espèce. Ensuite, nous étudions la compilation des espèces et collections vers le langage d'exécution OCAML, en utilisant les objets et modules OCAML. Puis nous détaillons la traduction dans le langage de preuves COQ, la liaison retardée étant traduite par des lambda-abstractions. Nous montrons ensuite comment utiliser cette technique pour optimiser les exécutables OCAML. Enfin, nous prouvons que les analyses faites par le compilateur ainsi que les techniques de traduction sont conforme à la formalisation des espèces faites auparavant en COQ.

Pour pouvoir raisonner sur nos programmes, une méthode est de directement les écrire dans un assistant de preuve. Utilisant la correspondance de Curry-Howard, les programmes et les preuves ne font alors qu’un. Pour ne pas nuire à la cohérence logique de l’assistant de preuve, le système est obligé de restreindre les programmes à ne pas avoir d’effets de bord. Cependant, les effets de bord sont omniprésents et essentiels dans la programmation. Différentes techniques telles que les monades ou les effets algébriques ont alors émergé pour les modéliser offrant ainsi un moyen d'écrire des programmes impératifs dans des langages purement fonctionnels. C'est donc assez naturellement que les résultats de décennies de recherches investies pour raisonner sur des programmes impératifs tentent d'être adaptés au raisonnement sur des programmes avec effets. Dans cette thèse, nous nous intéressons d'abord à l'utilisation de la logique de séparation pour raisonner sur des programmes avec effets implémentés dans un assistant de preuve. Nous étudions une approche consistant à décrire les comportements des effets à l'aide d'un transformateur de prédicats. Nous nous concentrons d'abord sur la fraîcheur, puis sur le traitement de paquets et le zéro-copie. Pour étudier notre approche, nous nous appuyons sur deux exemples concrets qui sont le module SimplExpr de CompCert et la bibliothèque de décodeur Nom. Pour finir, pour compiler les analyseurs de paquets produits vers C, nous proposons une méthode par raffinement supprimant les continuations introduites par l'utilisation d'une monade libre et effectuant quelques optimisations
One way to reason about our programs is to write them directly into a proof assistant. Using the Curry-Howard correspondence, programs and proofs are then one. In order not to undermine the logical consistency of the proof assistant, the system is forced to restrict the programs to have no side effects. However, side effects are ubiquitous and essential in programming. Different techniques such as monads or algebraic effects have emerged to model them, thus offering a way to write imperative programs in purely functional languages. It is therefore quite natural that the results of decades of research invested in reasoning about imperative programs try to be adapted to reasoning about programs with effects. In this thesis, we are first interested in the use of separation logic to reason about programs with effects implemented in a proof assistant. We study an approach to describe the behaviour of effects using a predicate transformer. We focus first on freshness, then on packet processing and zero-copy. To study our approach, we rely on two concrete examples which are the SimplExpr module of CompCert and the decoder library Nom. Finally, in order to compile the packet parsers produced to C, we propose a refinement method removing the continuations introduced by the use of a free monad and performing some optimizations

JavaScript est un langage de programmation maintenant très utilisé - y compris dans des domaines où la sécurité est importante. Il est donc important de permettre de vérifier la qualité des logiciels écrit en JavaScript. Cette thèse explore l'approche de la preuve formelle, visant à donner une preuve mathématique qu'un programme donné se comporte comme prévu. Pour construire cette preuve, nous utilisons un assistant de preuve tel que Coq - un programme de confiance permettant de vérifier nos preuves formelles. Pour pouvoir énoncer qu'un programme JavaScript se comporte correctement, nous avons tout d'abord besoin d'une sémantique du langage JavaScript. Cette thèse s'est donc inscrite dans le projet JSCert visant à produire une sémantique formelle pour le langage JavaScript. Devant la taille de la sémantique de JavaScript, il est important de savoir comment on peut lui faire confiance : une faute de frappe peut compromettre toute la sémantique. Pour pouvoir faire confiance à JSCert, nous nous sommes appuyés sur deux sources de confiance. D'une part, JSCert a été conçue pour être très similaire à la spécification officielle de JavaScript, le standard ECMAScript : ils utilisent les mêmes structures de donnée, et il est possible d'associer chaque règle de réduction dans JSCert à une ligne d'ECMAScript. D'autre part, nous avons défini et prouvé relativement à JSCert un interpréteur nommé JSRef. Nous avons aussi pu lancer JSRef sur les suites de test de JavaScript. La sémantique de JSCert n'est pas la première sémantique formelle pour le JavaScript, mais c'est la première à proposer deux manières distinctes pour relier la sémantique formelle au langage JavaScript : en ayant une sémantique très similaire à la spécification officielle, et en ayant testé cette sémantique pour la comparer aux autres interpréteurs. Plutôt que de prouver indépendamment que chaque programme JavaScript s'exécute comme prévu, nous analysons ses programmes par interprétation abstraite. Cela consiste à interpréter la sémantique d'un langage avec des domaines abstraits. Par exemple la valeur concrète 1 pourra être remplacée par la valeur abstraite +. L'interprétation abstraite se compose en deux étapes : d'abord une sémantique abstraite est construite et prouvée correcte vis à vis de sa sémantique concrète, puis des analyseurs sont construits selon cette sémantique abstraite. Nous ne nous intéresserons qu'à la première étape dans cette thèse. La sémantique de JSCert est immense - plus de huit cent règles de réduction. La construction d'une sémantique abstraite traditionnelle ne passent pas à l'échelle face à de telles tailles. Nous avons donc conçu une nouvelle manière de construire des sémantiques abstraites à partir de sémantiques concrètes. Notre méthode se base sur une analyse précise de la structure des règles de réduction et vise à minimiser l'effort de preuve. Nous avons appliqué cette méthode sur plusieurs langages. Dans le but d'appliquer notre approche sur JavaScript, nous avons construit un domaine basé sur la logique de séparation. Cette logique requiert de nombreuses adaptations pour pouvoir s'appliquer dans le cadre de l'interprétation abstraite. Cette thèse en étudie les interactions et propose une nouvelle approche pour les solutionner dans le cadre construit précédemment. Nos domaines, bien qu'assez simple par rapport au modèle mémoire de JavaScript, semblent permettre la preuve d'analyseurs déjà existant. Les contributions de cette thèse sont donc triples : une sémantique formelle de confiance pour le langage JavaScript, un formalisme générique pour construire des sémantiques abstraites, et un domaine non trivial pour ce formalisme
JavaScript is a trending programming language. It is not used in applications in which security may be an important issue. It thus becomes important to be able to control the quality of softwares written in JavaScript. This thesis explores a formal proof approach, which aims at giving a mathematical proof that a given program behaves as expected. To build this proof, we use proof assistants such as Coq—a trusted program enabling to check formal proofs. To state that a JavaScript program is behaving as expected, we first need a semantics of the JavaScript language. This thesis is thus part of the JSCert project, whose aim it to prove a formal semantics for JavaScript. Because of the size of JavaScript's semantics, it is crucial to know how it can be trusted: a typing mistake could compromise the whole semantics. To trust JSCert, we based ourselves on two trust sources. On one hand, JSCert has been designed to be the most similar it can be from the official JavaScript specification, the ECMAScript standard: they use the same data structures, and it is possible to relate each derivation rule in JSCert to a line of ECMAScript. On the other hand, we defined and proved correct with respect to JSCert an interpreter named JSRef. We have been able to run JSRef on JavaScript test suites. The JSCert semantics is not the first formal semantics of JavaScript, but it is the first to propose two distinct ways to relate the formal semantics to the JavaScript language: by having a semantics close to the official specification, and by testing this semantics and comparing it to other interpreters. Instead of independently proving that each JavaScript program behaves as expected, we chose to analyse programs using abstract interpretation. It consists of interpreting the semantics of a programming language with abstract domains. For instance, the concrete value 1 can be replaced by the abstract value +. Abstract interpretation is split into two steps : first, an abstract semantics is built and proven correct with respect to its concrete semantics, then, analysers are built from this abstract semantics. We only focus on the first step in this thesis. The JSCert semantics is huge - more than height hundred derivation rules. Building an abstract semantics using traditional techniques does not scale towards such sizes. We thus designed a new way to build abstract semantics from concrete semantics. Our approach is based on a careful analysis on the structure of derivation rules. It aims at minimising the proof effort needed to build an abstract semantics. We applied our method on several languages. With the goal of applying our approach to JavaScript, we built a domain based on separation logic. This logic require several adaptations to be able to apply in the context of abstract interpretation. This thesis precisely studies these interactions and introduces a new approach to solve them in our abstract interpretation framework. Our domains, although very simple compared to the memory model of JavaScript, seems to enable the proof of already existing analysers. This thesis has thus three main contributions : a trusted formal semantics for the JavaScript, a generic framework to build abstract semantics, and a non-trivial domain for this formalism