Tesis sobre el tema "Automatic software transformation"

Le noyau Linux est aujourd'hui présent dans tous les types de systèmes informatiques, des smartphones aux supercalculateurs, comprenant à la fois le matériel le plus récent et les systèmes "anciens". Cette diversité d'environnement a pour conséquence une base de code importante, d'une dizaine de millions de lignes de code, pour les pilotes matériels. Cependant par souci d'introduction de nouvelles fonctionnalités, ou pour des raisons de performance ou de sécurité, certaines interfaces de programmation (APIs) internes doivent être parfois revues, ce qui peut impliquer des changements pour des milliers de pilotes les utilisant.Cette thèse propose une nouvelle approche, Spinfer, permettant d'effectuer ces migrations d'utilisation d'APIs de manière automatique. Cette nouvelle approche, basée sur l'assemblage de motifs en suivant des contraintes de flot de contrôle, est capable d'apprendre à partir d'exemples, même imparfaits, des règles de transformations adaptées aux enjeux des migrations d'utilisations d'APIs dans le noyau Linux
The Linux kernel is present today in all kinds of computing environments, from smartphones to supercomputers, including both the latest hardware and "ancient" systems. This multiplicity of environments has come at the expense of a large code size, of approximately ten million lines of code, dedicated to device drivers. However, to add new functionalities, or for performance or security reasons, some internal Application Programming Interfaces (APIs) can be redesigned, triggering the need for changes of potentially thousands of drivers using them.This thesis proposes a novel approach, Spinfer, that can automatically perform these API usage updates. This new approach, based on pattern assembly constrained by control-flow relationships, can learn transformation rules from even imperfect examples. Learned rules are suitable for the challenges found in Linux kernel API usage updates

The Model Driven Architecture (MDA) allows moving the software development from the time consuming and error-prone level of writing program code to the next higher level of modeling. In order to gain benefit from this innovative technology, it is necessary to satisfy two requirements. These are first, the creation of compact, complete and correct platform independent models (PIM) and second, the development of a flexible and extensible model transformation framework taking into account frequent changes of the target platform. In this thesis a platform-based methodology is developed to create PIM by abstracting common modeling elements into a platform independent modeling library called Design Platform Model (DPM). The DPM contains OCL-based types for modeling primitive and collection types, a platform independent GUI toolkit as well as other common modeling elements, such as those for IO-operations. Furthermore, a DPM profile containing diverse domain specific and design pattern-based stereotypes is also developed to create PIM with high-level semantics. The behavior in PIM is specified using an OCL-like action language called eXecutable OCL (XOCL), which is also developed in this thesis. For model transformation, the model compiler MOCCA is developed based on a flexible and extensible architecture. The model mapper components in the current version of MOCCA are able to map desktop applications onto JSE platform; the both business object layer and persistence layer of a three-layered enterprise applications onto JEE platform and SAP ABAP platform. The entire model transformation process is finished with complete code generation.

Dissertação apresentada na Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa para a obtenção do Grau de Mestre em Engenharia Informática.
Many programs sequentially execute operations that take a long time to complete. Some of these operations may return a highly predictable result. If this is the case, speculative execution can improve the overall performance of the program. Speculative execution is the execution of code whose result may not be needed. Generally it is used as a performance optimization. Instead of waiting for the result of a costly operation,speculative execution can be used to speculate the operation most probable result and continue executing based in this speculation. If later the speculation is confirmed to be correct, time had been gained. Otherwise, if the speculation is incorrect, the execution based in the speculation must abort and re-execute with the correct result. In this dissertation we propose the design of an abstract process to add speculative execution to a program by doing source-to-source transformation. This abstract process is used in the definition of a mechanism and methodology that enable programmer to add speculative execution to the source code of programs. The abstract process is also used in the design of an automatic source-to-source transformation process that adds speculative execution to existing programs without user intervention. Finally, we also evaluate the performance impact of introducing speculative execution in database clients. Existing proposals for the design of mechanisms to add speculative execution lacked portability in favor of performance. Some were designed to be implemented at kernel or hardware level. The process and mechanisms we propose in this dissertation can add speculative execution to the source of program, independently of the kernel or hardware that is used. From our experiments we have concluded that database clients can improve their performance by using speculative execution. There is nothing in the system we propose that limits in the scope of database clients. Although this was the scope of the case study, we strongly believe that other programs can benefit from the proposed process and mechanisms for introduction of speculative execution.

The goals of this work are a design and implementation of a new vision system, integrated with the waterjet machine. This system combines two commercial webcams applied on an industrial dedicated platform. A main purpose of the vision system is to detect the position and rotation of a workpiece placed on the machine table. The used object recognition algorithm consists of edge detection, standard math processing functions and noise filters. The Hough transform technique is used to extract lines and their intersections of a workpiece. Metric rectification method is used, in order to obtain a top view of the workspace and to adjust an image coordinate system, accordingly to the waterjet machine coordinates. In situ calibration procedures of the booth webcams are developed and implemented. Experimental results of the proposed new vision system prototype confirm required performance and precision of the element detection.

Browsing the Web on a distant screen may very often become user’s unpleasant experience. Such a situation may happen in a living room environment, where the user attempts to browse the Web on TV sitting at 10-foot distance from the screen. In a vast number of web sites that the user is able to request, many may contain, for instance, tiny text characters, which would become unreadable when watched from the distance. On the other hand, also navigation through the Web offerings can become troublesome. Especially, when the user is only equipped with the commonly used remote controller, interaction with the displayed content may prove inefficient. Therefore, methods to overcome content visibility and navigation problems become challenging. In this thesis we perceive living room environment as a good scenario, in which web-enabled handheld devices could further be utilized to interact with Consumer Electronics products, such as TVs. This sort of interaction should make an advantage of publically available Web resources to have them displayed on TV and controlled from the distance by the user who is equipped with, for instance, a web-enabled mobile phone. We think that such an interaction approach would become beneficial in several ways, making the Web browsing routine an interactive, possibly a group activity. Discussed in this thesis solutions are the first step of automatic transformations of web content into a form that can be rendered on Web4CE (Web for Consumer Electronics) products. A particularly important aspect of the proposed system is the concept of request/response HTTP message processing and HTML modification. According to this approach, living room browsing activity should be achievable without the need to re-author already existing web sites.

Les microprocesseurs et les architectures parallèles d'aujourd'hui lancent de nouveaux défis aux techniques de compilation. En présence de parallélisme, les optimisations deviennent trop spécifiques et complexes pour être laissées au soin du programmeur. Les techniques de parallélisation automatique dépassent le cadre traditionnel des applications numériques et abordent de nouveaux modèles de programmes, tels que les nids de boucles non affines, les appels récursifs et les structures de données dynamiques. Des analyses précises sont au c{\oe}ur de la détection du parallélisme, elles rassemblent des informations à la compilation sur les propriétés des programmes à l'exécution. Ces informations valident des transformations utiles pour l'extraction du parallélisme et la génération de code parallèle. Cette thèse aborde principalement des analyses et des transformations avec une vision par instances, c'est-à-dire considérant les propriétés individuelles de chaque instance d'une instruction à l'exécution. Une nouvelle formalisation à l'aide de langages formels nous permet tout d'abord d'étudier une analyse de dépendances et de définitions visibles par instances pour programmes récursifs. L'application de cette analyse à l'expansion et la parallélisation de programmes récursifs dévoile des résultats encourageants. Les nids de boucles quelconques font l'objet de la deuxième partie de ce travail. Une nouvelle étude des techniques de parallélisation fondées sur l'expansion nous permet de proposer des solutions à des problèmes d'optimisation cruciaux.

This thesis is on the automatic creation of model generators to assist the validation of model transformations. The model driven software development methodology advocates models as the main artefact to represent software during development. Such models are automatically converted, by transformation tools, to apply in different stages of development. In one application of the method, it becomes possible to synthesise software implementations from design models. However, the transformations used to convert models are man-made, and so prone to development error. An error in a transformation can be transmitted to the created software, potentially creating many invalid systems. Evaluating that model transformations are reliable is fundamental to the success of modelling as a principle software development practice. Models generated via the technique presented in this thesis can be applied to validate transformations. In several existing transformation validation techniques, some form of conversion is employed. However, those techniques do not apply to validate the conversions used there-in. A defining feature of the current presentation is the utilization of transformations, making the technique self-hosting. That is, an implementation of the presented technique can create generators to assist model transformations validation and to assist validation of that implementation of the technique.

SIMON is an automated testing application developed by WM-Data Consulting in Växjö, Sweden. Previously the test cases, called BIFs, run by SIMON to test the applications under test has been written manually in a very time consuming manner offering no protection against errors in the structure or misspellings. This thesis investigates a replacement to the manual method when creating the BIFs; my own developed application called the BIF-Editor. The usage of the BIF-Editor guaranteed correct syntax and structure and made the creation of the BIFs faster, but did it increase the quality of the BIFs? So to evaluate the BIF-Editor, the quality regarding path coverage of BIFs manually created was compared with BIFs created during the same elapsed time using the BIF-Editor. This evaluation showed that the usage of the BIF-Editor increased the quality of the BIFs by making the creation safer, but primarily faster which enabled the user to produce more BIFs than previously possible resulting in a raised path cover.

This thesis presents an approach for working with low level source code that enables automatic restructuring and raising the abstraction level of the programs. This makes it easier to understand the logic of the program, which in turn reduces the development time.The process in this thesis was designed to be flexible and consists of several independent tools. This makes the process easy to adapt as needed, while at the same time the developed tools can be used for other processes. There are usually two basic steps. First is the translation to WSL language, which has a great number of semantic preserving program transformations. The second step are the transformations of the translated WSL. Two tools were developed for translation: one that works with a subset of x86 assembly, and another that works with MicroJava bytecode. The result of the translation is a low level program in WSL.The primary goal of this thesis was to fully automate the selection of the transformations. This enables users with no domain  knowledge to efficiently use this process as needed. At the same time, the flexibility of the process enables experienced users to adapt it as needed or integrate it into other processes. The automation was achieved with a hill climbing algorithm.Experiments that were run on several types of input programs showed that the results can be excellent. The fitness function used was a built-in metric that gives the “weight” of structures in a program. On input samples that had original high level source codes, the end result metrics of the translated and transformed programs were comparable. On some samples the result was even better than the originals, on some others they were somewhat more complex. When comparing with low level original source code, the end results was always significantly improved.
U okviru ove teze se predstavlja pristup radu sa programima niskog nivoa koji omogućava automatsko restrukturiranje i podizanje na više nivoe. Samim tim postaje mnogo lakše razumeti logiku programa što smanjuje vreme razvoja.Proces je dizajniran tako da bude fleksibilan i sastoji se od više nezavisnih alata. Samim tim je lako menjati proces po potrebi, ali i upotrebiti razvijene alate u drugim procesima. Tipično se mogu razlikovati dva glavna koraka. Prvi je prevođenje u jezik WSL,za koji postoji veliki broj transformacija programa koje očuvavaju semantiku. Drugi su transformacije u samom WSL-u. Za potrebe prevođenja su razvijena dva alata, jedan koji radi sa podskupom x86 asemblera i drugi koji radi sa MikroJava bajtkôdom. Rezultat prevođenja je program niskog nivoa u WSL jeziku.Primarni cilj ovog istraživanja je bila potpuna automatizacija odabira transformacija, tako da i korisnici bez iskustva u radu sa sistemom mogu efikasno da primene ovaj proces za svoje potrebe. Sa druge strane zbog fleksibilnosti procesa, iskusni korisnici mogu lakoda ga prošire ili da ga integrišu u neki drugi već postojeći   proces.Automatizacija je  postignuta pretraživanjem usponom (eng. hill climbing).Eksperimenti vršeni na nekoliko tipova ulaznih programa niskog nivoa su pokazali da rezultati mogu biti  izuzetni. Za funkciju pogodnosti je korišćena ugrađena metrika koja daje “težinu” struktura u programu. Kod ulaza za koje je originalni izvorni kôd bio dostupan, krajnje metrike najboljih varijanti prevedenih i transformisanih programa su bile na sličnom nivou. Neki primeri su bolji od originala, dok su drugi bili nešto kompleksniji. Rezultati su uvek pokazivali značajna unapređenja u odnosu na originalni kôd niskog nivoa.

De nombreux prouveurs et outils de vérification font un usage instensif des transformations logiques afin de ramener un problème exprimé sous la forme d'une tâche de preuve à un certain nombre de tâches de preuve plus simples à valider. Les transformations font souvent partie de la base de confiance de l'outil de vérification. Cette thèse a pour objectif de renforcer la confiance accordée aux transformations logiques. Les transformations sont instrumentées pour produire des certificats puis ceux-ci sont vérifiés par un outil externe: c'est l'approche sceptique. De ce fait, notre méthode est incrémentale et robuste aux modifications apportées au code des transformations. Nous définissons deux formats de certificats; les transformations génèrent des certificats de surface et ces certificats sont traduits en des certificats de noyau qui sont destinés à la vérification finale. L'accent est mis sur la facilité de production des certificats de surface et nous avons fait en sorte qu'ils soient, autant que possible, indépendants des tâches de preuve, facilitant ainsi leur composition et rendant la certification plus modulaire. Les certificats de noyau, au contraire, incluent de nombreux détails tout en restant élémentaires, de sorte que leur vérification est réalisable par un outil simple, dont la confiance est facile à établir. Nous proposons une procédure de traduction d'un certificat de surface en un certificat de noyau qui n'a pas besoin d'être certifiée. Les transformations logiques sont considérées dans une logique d'ordre supérieur avec polymorphisme de type, ce formalisme pouvant être étendu avec des théories interprétées telles que l'égalité ou l'arithmétique entière. Nous appliquons notre méthode à Why3, et notamment à des transformations complexes qui pré-existent à notre travail. Nous implémentons également deux vérificateurs de certificats, le premier reposant sur une approche calculatoire efficace et l'autre s'appuyant sur un encodage superficiel des tâches de preuve dans le framework logique Lambdapi, donnant ainsi de fortes garanties de sa correction
In various provers and deductive verification tools, logical transformations are used extensively in order to reduce a proof task into a number of simpler tasks. Logical transformations are often part of the trusted base of such tools. In this thesis, we develop a framework to improve confidence in their results. We follow a skeptical} approach: transformations are instrumented to produce certificates that are checked by a third-party tool. Thus, we benefit from a modular approach that is also robust to changes in the source code of the transformations. We design two kinds of certificates. Transformations produce surface certificates} that are then translated to kernel certificates} which are destined for final verification. We made sure that surface certificates are easy to produce. Moreover, surface certificates are as independent of the transformation application as possible and this makes for a modular certification. On the contrary, kernel certificates include numerous details about the transformation application and are kept elementary. This helps to define simpler checkers and establish their correctness. We propose a translation procedure from surface certificates to kernel certificates which does not need to be trusted. Logical transformations are considered in a higher-order logic, with type polymorphism and built-in theories such as equality and integer arithmetic. We apply our framework to Why3 and use it to instrument pre-existing and complex transformations. Additionally, we implement two certificate checkers. The first one follows an efficient computational approach while the second is based on a shallow embedding of proof tasks inside the logical framework Lambdapi, thus exhibiting formal guaranties of its correctness

Диссертация на соискание ученой степени доктора технических наук по специальности 05.13.07 – автоматизация процессов управления. – Национальный технический университет "Харьковский политехнический институт", Харьков, 2017. Диссертация посвящена решению научно-прикладной проблемы разработки бортовой системы поддержки принятия решений машинистом, созданной на основе обобщенных математических моделей и средств оптимизации динамики подвижных объектов с использованием новых методов и программного обеспечения, а также новой технологии обработки информации на основе стабильно-пластичных нейронных сетей и новых моделей ассоциативной памяти, что создает теоретическую предпосылку разработки автоматических систем управления подвижным составом и позволяет улучшить его энергетические характеристики. Разработана модель дизель-поезда, учитывающая основные виды колебаний вагонов и распределение сил взаимодействия между ними во время движения, а также параллельную работу тяговых приводов, которая адекватно отражает процессы, протекающие на реальном объекте. Разработано программное обеспечение, реализующее человеко-машинную систему, которая позволяет автоматизировать аналитические преобразования геометрической теории управления при синтезе моделей в форме Бруновского для объектов, описываемых системами обыкновенных дифференциальных уравнений высокого порядка с несколькими управлениями. На основе нейронных сетей адаптивной резонансной теории, способных решать задачи с несколькими решениями, предложен новый метод поиска функций преобразования между переменными линейных и нелинейных моделей. С помощью принципа максимума решены две задачи оптимального управления тяговым подвижным составом: максимального быстродействия и минимизации взвешенной линейной комбинации времени и расходов квадрата управления, что позволяет, с одной стороны, получить для каждого участка пути законы управления, которые определяют минимально необходимое время для преодоления перегона, а с другой стороны, получить законы управления, обеспечивающие график движения и минимизацию расхода топливно-энергетических ресурсов. Разработаны стабильно-пластичные нейронные сети Хемминга, Хебба и сети на основе перцептрона, способные распознавать новую информацию и дообучаться в процессе своего функционирования, которые позволяют использовать их как альтернативу дискретным нейронным сетям адаптивной резонансной теории. Разработана бортовая система поддержки принятия решений, которая позволяет в реальных условиях скоростного движения выдавать машинисту закон управления поездом, при котором соблюдается график движения при минимальных затратах топливно-энергетических ресурсов. Для реализации базы данных системы поддержки принятия решений машинистом разработана N-направленная нейросетевая ассоциативная память, которая способна восстанавливать по входному вектору множество из N векторов, ассоциативных к входной информации, и двунаправленная многослойная дискретная ассоциативная память с управляющими нейронами, которая способна восстанавливать цепочки ассоциаций и корректировать результаты с учетом дополнительной информации. Создана база знаний, позволяющая запоминать несколько равноценных решений о законе управления поездом для текущего перегона, нейросетевая система диагностики тяговых двигателей и система, позволяющая прогнозировать возникновение и подавлять развитие буксования колесных пар во время движения. Приведены результаты экспериментальных исследований интеллектуальной системы поддержки принятия решений машинистом и законов оптимального управления подвижным составом, которые подтвердили достоверность предложенных решений по автоматизации процессов управления движением дизель-поезда.
Dissertation for the degree of Doctor of Technical Science on Specialty 05.13.07 – Automation of Control Processes. – National Technical University "Kharkіv Polytechnic Institute", Kharkіv, 2017. The dissertation is devoted to solving scientific and applied problems of increasing energy characteristics of traction rolling stock, traction asynchronous drive through the development and introduction of new onboard intellectual decision support system of machinist that is based on more accurate models and optimize dynamics tools, using new methods and specialized software and new technologies of information processing based on stable and plastic neural networks and new models of associative memory, which creates theoretical precondition for developing of automatic control systems of high-speed rolling stock. Developed a comprehensive diesel-train model that takes into account the principal vibrations of the rolling stock and the distribution of forces of interaction between them, as well as parallel performance of traction drive of wagons with motors that adequately reflects the processes that occur on the real object. Developed specialized software that realizes the man-machine system that automates the conversion of analytical geometric control theory in the synthesis models to the Brunovsky form. Based on the neural networks that can solve problems with multiple solutions, developed a new method of search switching functions between the variables in the form of linear models Brunovsky and variables of nonlinear models control object. Developed stable-flexible Hamming neural networks, Hebb and other networks based on Perceptron, that can recognize new information and studying during its performance and modification Hamming neural network, capable identify several solutions. Has developed structure and components of the onboard intellectual decision support system that allows the actual use of the rolling stock and when happening currently changing road conditions, system can give for machinist new control laws under which adheres to a schedule for the least cost fuel and energy resources. Conducted experimental investigations on mathematical models and real object that confirming the correctness of the proposed solutions, methods and algorithms.

Дисертація на здобуття наукового ступеня доктора технічних наук за спеціальністю 05.13.07 – автоматизація процесів керування. – Національний технічний університет "Харківський політехнічний інститут", Харків, 2017. Дисертація присвячена вирішенню науково-прикладної проблеми розробки бортової системи підтримки прийняття рішень машиністом, що створена на основі узагальнених математичних моделей та засобів оптимізації динаміки рухомих об'єктів з використанням нових методів та програмного забезпечення, а також нової технології обробки інформації на основі стабільно-пластичних нейронних мереж та нових моделей асоціативної пам'яті, яка створює теоретичну передумову розробки автоматичних систем керування рухомим складом та дозволяє поліпшити його енергетичні характеристики. Розроблено модель дизель-поїзда, що враховує основні види коливань вагонів та розподіл сил взаємодії між ними, а також паралельну роботу тягових двигунів обмоторених вагонів, що адекватно відображає процеси, які протікають на реальному об'єкті. Розроблено спеціалізоване програмне забезпечення, що реалізує людино-машинну систему, яка дозволяє автоматизувати аналітичні перетворення геометричної теорії керування при синтезі моделей у формі Бруновського. На основі нейронних мереж, що здатні вирішувати завдання з декількома рішеннями, розроблено новий метод пошуку функцій переходу між змінними нелінійних і лінійних моделей у формі Бруновського. Розроблені стабільно-пластичні нейронні мережі Хеммінга, Хебба та мережі на основі перцептрона, здатні розпізнавати нову інформацію й донавчатися в процесі функціонування, та модифікацію нейронної мережі Хеммінга, що здатна визначати декілька рішень. Розроблено структуру та складові бортової системи підтримки прийняття рішень, що дозволяє в реальних умовах експлуатації рухомого складу й поточній зміні дорожньої обстановки видавати машиністу закон керування, при якому дотримується графік руху за мінімальних витрат паливо-енергетичних ресурсів. Проведені дослідження на математичних моделях і реальному об'єкті, результати яких підтверджують правильність запропонованих рішень, методів та алгоритмів.
Dissertation for the degree of Doctor of Technical Science on Specialty 05.13.07 – Automation of Control Processes. – National Technical University "Kharkіv Polytechnic Institute", Kharkіv, 2017. The dissertation is devoted to solving scientific and applied problems of increasing energy characteristics of traction rolling stock, traction asynchronous drive through the development and introduction of new onboard intellectual decision support system of machinist that is based on more accurate models and optimize dynamics tools, using new methods and specialized software and new technologies of information processing based on stable and plastic neural networks and new models of associative memory, which creates theoretical precondition for developing of automatic control systems of high-speed rolling stock. Developed a comprehensive diesel-train model that takes into account the principal vibrations of the rolling stock and the distribution of forces of interaction between them, as well as parallel performance of traction drive of wagons with motors that adequately reflects the processes that occur on the real object. Developed specialized software that realizes the man-machine system that automates the conversion of analytical geometric control theory in the synthesis models to the Brunovsky form. Based on the neural networks that can solve problems with multiple solutions, developed a new method of search switching functions between the variables in the form of linear models Brunovsky and variables of nonlinear models control object. Developed stable-flexible Hamming neural networks, Hebb and other networks based on Perceptron, that can recognize new information and studying during its performance and modification Hamming neural network, capable identify several solutions. Has developed structure and components of the onboard intellectual decision support system that allows the actual use of the rolling stock and when happening currently changing road conditions, system can give for machinist new control laws under which adheres to a schedule for the least cost fuel and energy resources. Conducted experimental investigations on mathematical models and real object that confirming the correctness of the proposed solutions, methods and algorithms.

The Model Driven Architecture (MDA) allows moving the software development from the time consuming and error-prone level of writing program code to the next higher level of modeling. In order to gain benefit from this innovative technology, it is necessary to satisfy two requirements. These are first, the creation of compact, complete and correct platform independent models (PIM) and second, the development of a flexible and extensible model transformation framework taking into account frequent changes of the target platform. In this thesis a platform-based methodology is developed to create PIM by abstracting common modeling elements into a platform independent modeling library called Design Platform Model (DPM). The DPM contains OCL-based types for modeling primitive and collection types, a platform independent GUI toolkit as well as other common modeling elements, such as those for IO-operations. Furthermore, a DPM profile containing diverse domain specific and design pattern-based stereotypes is also developed to create PIM with high-level semantics. The behavior in PIM is specified using an OCL-like action language called eXecutable OCL (XOCL), which is also developed in this thesis. For model transformation, the model compiler MOCCA is developed based on a flexible and extensible architecture. The model mapper components in the current version of MOCCA are able to map desktop applications onto JSE platform; the both business object layer and persistence layer of a three-layered enterprise applications onto JEE platform and SAP ABAP platform. The entire model transformation process is finished with complete code generation.

Business processes describe a set of tasks for accomplishing business objectives of an organization. Business applications automate business processes to improve the productivity of business users. Nowadays, the business environment changes fast due to rapid market growth and technological innovations. Business processes are continuously updated to reflect new business initiatives. Business applications are frequently evolved to add features to meet new requirements and fix defects. Quite often, business processes and business applications evolve independently without direct reference to each other. Over time, it becomes more and more challenging to maintain the consistency between a business application and the corresponding business processes. Moreover, the existing development approaches rely on software developers’ craftsmanship to design and implement business applications. Such a development paradigm is inefficient and leads to inconsistency between business processes and business applications. To facilitate the alignment between business applications and business processes, we present an approach that automatically generates software architecture and code skeletons of business applications from business processes. We identify architectural components from business processes by analyzing dependencies among tasks. To verify the achievement of quality requirements, we extend a set of existing product metrics to automatically evaluate the quality of the generated software architecture designs. Eventually, we apply refactoring strategies, such as software architectural styles or design patterns, to optimize the generated software architecture designs and resolve identified quality problems. Moreover, we also present an approach to automatically refine software architecture to design models and code skeletons of business applications. The effectiveness of our proposed approach is illustrated through case studies.
Thesis (Ph.D, Electrical & Computer Engineering) -- Queen's University, 2011-01-30 00:06:34.77

Expert design knowledge is essential to develop a library of high-performance software. This includes how to implement and parallelize domain operations, how to optimize implementations, and estimates of which implementation choices are best. An expert repeatedly applies his knowledge, often in a rote and tedious way, to develop all of the related functionality expected from a domain-specific library. Expert knowledge is hard to gain and is easily lost over time when an expert forgets or when a new engineer starts developing code. The domain of dense linear algebra (DLA) is a prime example with software that is so well designed that much of experts' important work has become tediously rote in many ways. In this dissertation, we demonstrate how one can encode design knowledge for DLA so it can be automatically applied to generate code as an expert would or to generate better code. Further, the knowledge is encoded for perpetuity, so it can be reused to make implementing functionality on new hardware easier or it can be used to teach how software is designed to a non-expert. We call this approach to software engineering (encoding expert knowledge and automatically applying it) Design by Transformation (DxT). We present our vision, the methodology, a prototype code generation system, and possibilities when applying DxT to the domain of dense linear algebra.
text

Preemptive Time Petri Nets (pTPNs) support modeling and analysis of concurrent timed software components running under fixed priority preemptive scheduling. The model is supported by a well established theory based on symbolic state-space analysis through Difference Bounds Matrix (DBM), with specific contributions on compositional modularization, trace analysis, and efficient over-approximation and clean-up in the management of suspension deriving from preemptive behavior. The aim of this dissertation is to devise and implement a framework that brings the theory to application. To this end, the theory is cast into an organic tailoring of design, coding, and testing activities within a V-Model software life cycle in respect of the principles of regulatory standards applied to the construction of safety-critical software components. To implement the toolchain subtended by the overall approach into a Model Driven Development (MDD) framework, the theory of state-space analysis is complemented with methods and techniques supporting semi-formal specification and automated compilation into pTPN models and real-time code, measurement-based Execution Time estimation, test-case selection and sensitization, coverage evaluation.