No description available.

A mature code-base of 1 300 000 LOC for a period of 20 months has been examined. This paper investigates if churn is a significant factor in finding refactoring candidates. In addition it looks at the variables Lines of Code (LOC), Technical Debt (TD), Duplicated lines and Complexity to find out if any of these indicators can inform a coder as to what to refactor. The result is that churn is the strongest variable out of the studied variables followed by LOC and TD. / En kodbas på 1 300 000 rader kod har undersökts under 20 månader. Denna uppsats undersöker om kodens användningsfrekvens är en signifikant faktor för att finna refaktoreringskandidater. Uppsatsen tittar även antal kodrader, teknisk skuld, antal duplicerade kodrader och komplexitet för att undersöka om dessa indikatorer kan informera en programmerare om vad som ska refaktoreras. Resultatet är att kodens användningsfrekvens är den starkaste variabeln följt av antal kodrader samt teknisk skuld.

Le phénomène de l'alternance codique a été connu et étudié depuis le début du XXe siècle. L'intérêt psycholinguistique à l'alternance codique orale a été suivi par des perspectives socio-pragmatiques dans la seconde moitié du XXe siècle ; puis vers la fin du siècle, l'attention des chercheurs a été attirée sur l'alternance codique écrite. Même si l'alternance codique a été bien étudié depuis, il y a encore quelques lacunes. Par exemple, l'intersection de l'écriture bilingue et la critique génétique présente une source de données naturalistes bilingues écrites qui restent fortement peu étudiées. Cette thèse tente de combler cette lacune en fournissant une analyse qualitative et quantitative des brouillons d'examen des apprenants de langue seconde. L'approche multimodale englobe la perspective visuelle, pragmatique et linguistique. Les résultats montrent que l'alternance codique représente un outil que les élèves utilisent habilement dans leur écriture personnelle académique. Néanmoins, les objectifs de son utilisation diffèrent non seulement entre les individus mais aussi entre les sociétés / communautés. Cela souligne la valeur sociolinguistique de l'écriture personnelle bilingue. Ce travail démontre également l'alternance codique dans l'écriture personnelle comme un type indépendant de l'alternance codique, qui peut prendre une place égale à côté de l'alternance codique orale, écrite et littéraire et qui mérite d'être étudié en tant que tel. / The phenomenon of code-switching has been known and researched since the beginning of the 20th century. The psycholinguistic interest in spoken code-switching was followed by socio-pragmatic perspectives in the second half of the 20th century; then towards the end of the century, researchers' attention was drawn to code-switching in writing. Even though language alternation has been well studied since, there are still some lacunae. For example, the intersection of bilingual writing and genetic criticism presents a source of naturalistic bilingual written data which remains significantly under-studied. The present thesis attempts to fill this gap by providing a qualitative and quantitative analysis of second-language learners' examination drafts. The multimodal approach encompasses the visual, the pragmatic and the linguistic perspectives. The findings show that code-switching represents a tool which students skillfully use in their academic personal writing. However, the purposes of its use differ not only among individuals but also among societies/communities. This emphasizes the sociolinguistic value of bilingual personal writing. This work also recognizes code-switching in personal writing as an independent type of code-switching, which can be placed in line with the oral, written and literary code-switching and which merits to be studied as such.

The need for automated methods for identifying refactoring items is prelevent in many software projects today. Symptoms of refactoring needs is the concept of code smells within a software system. Recent studies have used single model machine learning to combat this issue. This study aims to test the possibility of improving machine learning code smell detection using ensemble methods. Therefore identifying the strongest ensemble model in the context of code smells and the relative sensitivity of the strongest perfoming ensemble identified. The ensemble models performance was studied by performing experiments using WekaNose to create datasets of code smells and Weka to train and test the models on the dataset. The datasets created was based on Qualitas Corpus curated java project. Each tested ensemble method was then compared to all the other ensembles, using f-measure, accuracy and AUC ROC scores. The tested ensemble methods were stacking, voting, bagging and boosting. The models to implement the ensemble methods with were models that previous studies had identified as strongest performer for code smell identification. The models where Jrip, J48, Naive Bayes and SMO. The findings showed, that compared to previous studies, bagging J48 improved results by 0.5%. And that the nominally implemented baggin of J48 in Weka follows best practices and the model where impacted negatively. However, due to the complexity of stacking and voting ensembles further work is needed regarding stacking and voting ensemble models in the context of code smell identification.

Software developers, maintainers, and testers rely on documentation to understandthe code they are working with. However, software documentation is perceivedas a waste of effort because it is usually outdated. How documentation evolvesthrough a set of releases may show whether there is any relationship between timeand quality. The results could help future developers and managers to improvethe quality of their documentation and decrease the time developers use to analyzecode. Previous studies showed that documentation used to be scarce and low inquality, thus, this research has investigated different variables to check if the qualityof the documentation changes over time. Therefore, we have created a tool thatwould extract and calculate the quality of the comments in code blocks, classes,and methods. The results have agreed with the previous studies. The quality of thedocumentation is affected to some extent through the releases, with a tendency todecrease.

L’utilisation des codes MDPC (Moderate Density Parity Check) quasi-cycliques dans le cryptosystème de McEliece offre un schéma de chiffrement post-quantique dont les clés ont une taille raisonnable et dont le chiffrement et le déchiffrement n’utilisent que des opérations binaires. C’est donc un bon candidat pour l’implémentation embarquée ou à bas coût.Dans ce contexte, certaines informations peuvent être exploitées pour construire des attaques par canaux cachés.Ici, le déchiffrement consiste principalement à décoder un mot de code bruité. Le décodeur utilisé est itératif et probabiliste : le nombre d’itérations de l'algorithme varie en fonction des instances et certains décodages peuvent échouer. Ces comportements ne sont pas souhaitables car ils peuvent permettre d’extraire des informations sur le secret.Une contremesure possible est de limiter le nombre d’instances de chiffrement avec les mêmes clés. Une autre façon serait de recourir à un décodage à temps constant dont la probabilité d’échec au décodage est négligeable. L’enjeu principal de cette thèse est de fournir de nouveaux outils pour analyser du comportement du décodeur pour la cryptographie.Dans un second temps, nous expliquons pourquoi l'utilisation des codes polaires n'est pas sûre pour le cryptosystème de McEliece. Pour ce faire, nous utilisons de nouvelles techniques afin de résoudre une équivalence de codes. Nous exhibons de nombreux liens entre les codes polaires et les codes de Reed-Muller et ainsi d'introduire une nouvelle famille de codes : les codes monomiaux décroissants. Ces résultats sont donc aussi d'un intérêt indépendant pour la théorie des codes. / Considering the McEliece cryptosystem using quasi-cylcic MDPC (Moderate Density Parity Check matrix) codes allows us to build a post-quantum encryption scheme with nice features. Namely, it has reasonable key sizes and both encryption and decryption are performed using binary operations. Thus, this scheme seems to be a good candidate for embedded and lightweight implementations. In this case, any information obtained through side channels can lead to an attack. In the McEliece cryptosystem, the decryption process essentially consists in decoding. As we consider the use of an iterative and probabilistic algorithm, the number of iterations needed to decode depends on the instance considered and some of it may fail to be decoded. These behaviors are not suitable because they may be used to extract information about the secrets. One countermeasure could be to bound the number of encryptions using the same key. Another solution could be to employ a constant time decoder with a negligible decoding failure probability, that is to say which is about the expected security level of the cryptosystem. The main goal of this thesis is to present new methods to analyse decoder behavior in a cryptographic context.Second, we explain why a McEliece encryption scheme based on polar code does not ensure the expected level of security. To do so, we apply new techniques to resolve the code equivalence problem. This allows us to highlight several common properties shared by Reed-Muller codes and polar codes. We introduce a new family of codes, named decreasing monomial codes, containing both Reed-Muller and polar codes. These results are also of independent interest for coding theory.

This research compares the energy performance of a code-built residential building within the moist climate zone classification in Canada, Europe, and the Northeastern United States. The primary objectives are to reveal how specific differences in code requirements in similar climates influence a building’s energy profile, offer a means to quantify and evaluate the extent of energy savings as a result of each requirement, and provide a comparison of each location’s building culture and how this affects the standards in place. Using the building energy simulation tool, DesignBuilder EnergyPlus Simulation, a model single-family home was created and input energy code requirements for each location. An evaluation of each location’s building culture is examined through such factors as the training of building professionals, commonly used materials and products, energy reduction goals, and cultural attitudes. The results of this study point to the need for more advanced building practices, stricter code mandates, and higher performing products based on energy savings achieved from buildings built to different standards in equivalent climate zones. This has the potential to drive the development and use of better performing building materials and assemblies in the future.

The standard genetic code (SGC) is the cipher used by nearly all organisms to transcribe information stored in DNA and translate it into its amino acid counterparts. Since the early 1960s, researchers have observed that the SGC is structured so that similar codons encode amino acids with similar physiochemical properties. This structure has been hypothesized to buffer the SGC against transcription or translational error because single nucleotide mutations usually either are silent or impart minimal effect on the containing protein. We herein briefly review different theories for the origin of that structure. We also briefly review different computational experiments designed to quantify buffering capacity for the SGC. We report on computational Monte Carlo simulations that we performed using a computer program that we developed, AGCT. In the simulations, the SGC was ranked against other, hypothetical genetic codes (HGC) for its ability to minimize physiochemical distances between amino acids encoded by codons separated by single nucleotide mutations. We analyzed unappreciated structural aspects and neglected properties in the SGC. We found that error measure type affected SGC ranking. We also found that altering stop codon positions had no effect on SGC ranking, but including stop codons in error calculations improved SGC ranking. We analyzed 49 properties individually and identified conserved properties. Among these, we found that long-range non-bonded energy is more conserved than is polar requirement, which previously was considered to be the most conserved property in the SGC. We also analyzed properties in combinations. We hypothesized that the SGC is organized as a compromise among multiple properties. Finally, we used AGCT to test whether different theories on the origin of the SGC could explain more convincingly the buffering capacity in the SGC. We found that, without accounting for transition/transversion biases, the SGC ranking was modest enough under constraints imposed by the coevolution and four column theories that it could be explained due to constraints associated with either theory (or both theories); however, when transition/transversion biases were included, only the four column theory returned a SGC ranking modest enough that it could be explained due to constraints associated with that theory. / Thesis / Master of Science (MSc) / The standard genetic code (SGC) is the cipher used almost universally to transcribe information stored in DNA and translate it to amino acid counterparts. Since the mid 1960s, researchers have recognized that the SGC is organized so that similar three-nucleotide RNA codons encode amino acids with similar properties; researchers consequently hypothesized that the SGC is structured to minimize effects from transcription or translation errors. This hypothesis has been tested using computer simulation. I briefly review results from those studies, complement them by analyzing unappreciated structural aspects and neglected properties, and test two theories on the origin of the SGC.

Developers build an entire software by writing quality code, the failure of which can risk organizational reputation, financial wellbeing, and even the lives and security of people. The ability to develop high quality code is therefore a key learning goal of foundational undergraduate computing programmes. However, how effective these undergraduate computing curriculums are in teaching software code quality is a matter of concern as this leads to raising the knowledge of future software developers. Right now, there is a lack of models to evaluate this effectiveness of undergraduate curriculum in providing the knowledge of software code quality to students.In this thesis, we suggest a model for evaluating the effectiveness of undergraduate curriculum in teaching software code quality, which we call ScQUc (Software Code Quality in Undergraduate Curriculum). Our goal is to raise the knowledge and awareness of teaching software code quality in education and make the undergraduate computing curriculum more software code quality centric.Due to the fact that there is a lack of models for evaluating the effectiveness of undergraduate curriculum in teaching software code quality, we had no closely related work to base our research on. We had to rely on other works related to code quality in programming courses that were indirectly connected to our research field. Hence, we dare infer that the ScQUc model was created from scratch.The research method used was qualitative and followed the frame of design science. Data collection was conducted via literature study and via interviews with two experts in the field. The interviewees were selected with the convenience sampling method enhanced with a predefined selection criterion. Data evaluation, on the other hand, was conducted using an evaluation model. The evaluation model included the following evaluation criteria: (1) interviewee credibility, (2) correctness, (3) flexibility, (4) usefulness and (5) experience. One round of interview was conducted with each of the interviewees which resulted in the improvement of the ScQUc model.The ScQUc model consists of a checklist of code quality characteristics and a sequence of activities to be conducted in order to evaluate the effectiveness of an undergraduate curriculum in teaching software code quality. The results of the evaluation of the ScQUc model inferred that the ScQUc model was correct and useful for the academia. Hence, we believe that the ScQUc model is a valid solution for evaluating the effectiveness of undergraduate curriculum in teaching software code quality. Still, however, the ScQUc model needs further evaluation and extension in form of specifications aiding deeper understanding of code quality, how to teach code quality, and to what extent to teach code quality. / Utvecklare bygger hela programvaran genom att skriva kvalitetskod, vars fel kan riskera organisatoriskt rykte, ekonomiskt välbefinnande och till och med människors liv och säkerhet. Förmågan att utveckla kod av hög kvalitet är därför ett viktigt inlärningsmål för grundläggande grundutbildningsprogram. Hur effektiva dessa läroplaner för grundutbildning är för att undervisa programkodkvalitet är emellertid en oro eftersom det leder till att kunskapen hos framtida programutvecklare ökar. Just nu saknas modeller för att utvärdera denna effektivitet i grundutbildningen för att ge studenterna kunskap om mjukvarukodkvalitet.I den här avhandlingen föreslår vi en modell för att utvärdera effektiviteten för grundutbildningen i undervisning av programkodkvalitet, som vi kallar ScQUc (Software Code Quality in Undergraduate Curriculum). Vårt mål är att öka kunskapen och medvetenheten om att undervisa programvarukodkvalitet i utbildningen och göra grundutbildningsberäkningen mer programvarukodkvalitet centrerad.På grund av bristen på modeller för att utvärdera effektiviteten i grundutbildningen i undervisning av programvarukodkvalitet, hade vi inget nära besläktat arbete att basera vår forskning på. Vi var tvungna att lita på andra verk relaterade till kodkvalitet i programmeringskurser som indirekt var kopplade till vårt forskningsområde. Därför drar vi slutsatsen att ScQUc-modellen skapades från grunden.Den använda metoden var kvalitativ och följde ramverket för designvetenskap. Datainsamlingen genomfördes genom litteraturstudie och genom intervjuer med två experter på området. Intervjuerna valdes med hjälp av metoden för provtagning av bekvämlighet förbättrad med ett fördefinierat urvalskriterium. Datautvärdering genomfördes å andra sidan med hjälp av en utvärderingsmodell. Utvärderingsmodellen inkluderade följande utvärderingskriterier: (1) intervjuarens trovärdighet, (2) korrekthet, (3) flexibilitet, (4) användbarhet och (5) erfarenhet. En intervjuomgång genomfördes med var och en av de intervjuade som resulterade i förbättringen av ScQUc-modellen.ScQUc-modellen består av en checklista med egenskaper för kodkvalitet och en sekvens av aktiviteter som ska genomföras för att utvärdera effektiviteten hos en grundutbildningsprogram i undervisning av programkodkvalitet. Resultaten av utvärderingen av ScQUc-modellen drar slutsatsen att ScQUc-modellen var korrekt och användbar för akademin. Därför tror vi att ScQUc-modellen är en giltig lösning för att utvärdera effektiviteten för grundutbildningen i undervisning av programvarukodkvalitet. Men ändå behöver ScQUc-modellen ytterligare utvärdering och utvidgning i form av specifikationer som hjälper djupare förståelse för kodkvalitet, hur man undervisar i kodkvalitet och i vilken utsträckning man ska lära ut kodkvalitet.

No description available.