A key collaborative hub for many software development projects is the issue tracking system, or bug repository. The use of a bug repository can improve the software development process in a number of ways including allowing developers who are geographically distributed to communicate about project development. However, reports added to the repository need to be triaged by a human, called the triager, to determine if reports are meaningful. If a report is meaningful, the triager decides how to organize the report for integration into the project's development process. We call triager decisions with the goal of determining if a report is meaningful, repository-oriented decisions, and triager decisions that organize reports for the development process, development-oriented decisions. Triagers can become overwhelmed by the number of reports added to the repository. Time spent triaging also typically diverts valuable resources away from the improvement of the product to the managing of the development process. To assist triagers, this dissertation presents a machine learning approach to create recommenders that assist with a variety of development-oriented decisions. In this way, we strive to reduce human involvement in triage by moving the triager's role from having to gather information to make a decision to that of confirming a suggestion. This dissertation introduces a triage-assisting recommender creation process that can create a variety of different development-oriented decision recommenders for a range of projects. The recommenders created with this approach are accurate: recommenders for which developer to assign a report have a precision of 70% to 98% over five open source projects, recommenders for which product component the report is for have a recall of 72% to 92%, and recommenders for who to add to the cc: list of a report that have a recall of 46% to 72%. We have evaluated recommenders created with our triage-assisting recommender creation process using both an analytic evaluation and a field study. In addition, we present in this dissertation an approach to assist project members to specify the project-specific values for the triage-assisting recommender creation process, and show that such recommenders can be created with a subset of the repository data.

A key collaborative hub for many software development projects is the issue tracking system, or bug repository. The use of a bug repository can improve the software development process in a number of ways including allowing developers who are geographically distributed to communicate about project development. However, reports added to the repository need to be triaged by a human, called the triager, to determine if reports are meaningful. If a report is meaningful, the triager decides how to organize the report for integration into the project's development process. We call triager decisions with the goal of determining if a report is meaningful, repository-oriented decisions, and triager decisions that organize reports for the development process, development-oriented decisions. Triagers can become overwhelmed by the number of reports added to the repository. Time spent triaging also typically diverts valuable resources away from the improvement of the product to the managing of the development process. To assist triagers, this dissertation presents a machine learning approach to create recommenders that assist with a variety of development-oriented decisions. In this way, we strive to reduce human involvement in triage by moving the triager's role from having to gather information to make a decision to that of confirming a suggestion. This dissertation introduces a triage-assisting recommender creation process that can create a variety of different development-oriented decision recommenders for a range of projects. The recommenders created with this approach are accurate: recommenders for which developer to assign a report have a precision of 70% to 98% over five open source projects, recommenders for which product component the report is for have a recall of 72% to 92%, and recommenders for who to add to the cc: list of a report that have a recall of 46% to 72%. We have evaluated recommenders created with our triage-assisting recommender creation process using both an analytic evaluation and a field study. In addition, we present in this dissertation an approach to assist project members to specify the project-specific values for the triage-assisting recommender creation process, and show that such recommenders can be created with a subset of the repository data.

A key collaborative hub for many software development projects is the issue tracking system, or bug repository. The use of a bug repository can improve the software development process in a number of ways including allowing developers who are geographically distributed to communicate about project development. However, reports added to the repository need to be triaged by a human, called the triager, to determine if reports are meaningful. If a report is meaningful, the triager decides how to organize the report for integration into the project's development process. We call triager decisions with the goal of determining if a report is meaningful, repository-oriented decisions, and triager decisions that organize reports for the development process, development-oriented decisions. Triagers can become overwhelmed by the number of reports added to the repository. Time spent triaging also typically diverts valuable resources away from the improvement of the product to the managing of the development process. To assist triagers, this dissertation presents a machine learning approach to create recommenders that assist with a variety of development-oriented decisions. In this way, we strive to reduce human involvement in triage by moving the triager's role from having to gather information to make a decision to that of confirming a suggestion. This dissertation introduces a triage-assisting recommender creation process that can create a variety of different development-oriented decision recommenders for a range of projects. The recommenders created with this approach are accurate: recommenders for which developer to assign a report have a precision of 70% to 98% over five open source projects, recommenders for which product component the report is for have a recall of 72% to 92%, and recommenders for who to add to the cc: list of a report that have a recall of 46% to 72%. We have evaluated recommenders created with our triage-assisting recommender creation process using both an analytic evaluation and a field study. In addition, we present in this dissertation an approach to assist project members to specify the project-specific values for the triage-assisting recommender creation process, and show that such recommenders can be created with a subset of the repository data. / Science, Faculty of / Computer Science, Department of / Graduate

Made available in DSpace on 2014-06-12T15:53:57Z (GMT). No. of bitstreams: 2 arquivo1938_1.pdf: 2696606 bytes, checksum: c2ff3cbbb3029fd0f89eb8d67c0e4f08 (MD5) license.txt: 1748 bytes, checksum: 8a4605be74aa9ea9d79846c1fba20a33 (MD5) Previous issue date: 2009 / Manutenção e evolução de software são atividades caracterizadas pelo seu enorme custo e baixa velocidade de execução. Não obstante, elas são atividades inevitáveis para garantir a qualidade do software quase todo software bem sucedido estimula os usuários a fazer pedidos de mudanças e melhorias. Sommerville é ainda mais enfático e diz que mudanças em projetos de software são um fato. Além disso, diferentes estudos têm afirmado ao longo dos anos que as atividades de manutenção e evolução de software são as mais caras do ciclo de desenvolvimento, sendo responsável por cerca de até 90% dos custos. Todas essas peculiaridades da fase de manutenção e evolução de software leva o mundo acadêmico e industrial a investigar constantemente novas soluções para reduzir os custos dessas atividades. Neste contexto, Gerência de Configuração de Software (GCS) é um conjunto de atividades e normas para a gestão da evolução e manutenção de software; GCS define como são registradas e processadas todas as modificações, o impacto das mesmas em todo o sistema, dentre outros procedimentos. Para todas estas tarefas de GCM existem diferentes ferramentas de auxílio, tais como sistemas de controle de versão e bug trackers. No entanto, alguns problemas podem surgir devido ao uso das mesmas, como por exemplo o problema de atribuição automática de responsável por um bug report e o problema de duplicação de bug reports. Neste sentido, esta dissertação investiga o problema de duplicação de bug reports resultante da utilização de bug trackers em projetos de desenvolvimento de software. Tal problema é caracterizado pela submissão de dois ou mais bug reports que descrevem o mesmo problema referente a um software, tendo como principais conseqüências a sobrecarga de trabalho na busca e análise de bug reports, e o mal aproveitamento do tempo destinado a essa atividade

A software Bug report contains information about the bug in the form of problem description and comments using natural language texts. Managing reported bugs is a significant challenge for a project manager when the number of bugs for a software project is large. Prior to the assignment of a newly reported bug to an appropriate developer, the triager (e.g., manager) attempts to categorize it into existing categories and looks for duplicate bugs. The goal is to reuse existing knowledge to fix or resolve the new bug, and she often spends a lot of time in reading a number of bug reports. When fixing or resolving a bug, a developer also consults with a series of relevant bug reports from the repository in order to maximize the knowledge required for the fixation. It is also preferable that developers new to a project first familiarize themselves with the project along with the reported bugs before actually working on the project. Because of the sheer numbers and size of the bug reports, manually analyzing a collection of bug reports is time-consuming and ineffective. One of the ways to mitigate the problem is to analyze summaries of the bug reports instead of analyzing full bug reports, and there have been a number of summarization techniques proposed in the literature. Most of these techniques generate extractive summaries of bug reports. However, it is not clear how useful those generated extractive summaries are, in particular when the developers do not have prior knowledge of the bug reports. In order to better understand the usefulness of the bug report summaries, in this thesis, we first reimplement a state of the art unsupervised summarization technique and evaluate it with a user study with nine participants. Although in our study, 70% of the time participants marked our developed summaries as a reliable means of comprehending the software bugs, the study also reports a practical problem with extractive summaries. An extractive summary is often created by choosing a certain number of statements from the bug report. The statements are extracted out of their contexts, and thus often lose their consistency, which makes it hard for a manager or a developer to comprehend the reported bug from the extractive summary. Based on the findings from the user study and in order to further assist the managers as well as the developers, we thus propose an interactive visualization for the bug reports that visualizes not only the extractive summaries but also the topic evolution of the bug reports. Topic evolution refers to the evolution of technical topics discussed in the bug reports of a software system over a certain time period. Our visualization technique interactively visualizes such information which can help in different project management activities. Our proposed visualization also highlights the summary statements within their contexts in the original report for easier comprehension of the reported bug. In order to validate the applicability of our proposed visualization technique, we implement the technique as a standalone tool, and conduct both a case study with 3914 bug reports and a user study with six participants. The experiments in the case study show that our topic analysis can reveal useful keywords or other insightful information about the bug reports for aiding the managers or triagers in different management activities. The findings from the user study also show that our proposed visualization technique is highly promising for easier comprehension of the bug reports.

Many bugs, even those that are known and documented in bug reports, remain in mature software for a long time due to the lack of the development resources to fix them. We propose a general approach, R2Fix, to automatically generate bug-fixing patches from free-form bug reports. R2Fix combines past fix patterns, machine learning techniques, and semantic patch generation techniques to fix bugs automatically. We evaluate R2Fix on three large and popular software projects, i.e., the Linux kernel, Mozilla, and Apache, for three important types of bugs: buffer overflows, null pointer bugs, and memory leaks. R2Fix generates 60 patches correctly, 5 of which are new patches for bugs that have not been fixed by developers yet. We reported all 5 new patches to the developers; 4 have already been accepted and committed to the code repositories. The 60 correct patches generated by R2Fix could have shortened and saved an average of 68 days of bug diagnosis and patch generation time.

Many bugs, even those that are known and documented in bug reports, remain in mature software for a long time due to the lack of the development resources to fix them. We propose a general approach, R2Fix, to automatically generate bug-fixing patches from free-form bug reports. R2Fix combines past fix patterns, machine learning techniques, and semantic patch generation techniques to fix bugs automatically. We evaluate R2Fix on three large and popular software projects, i.e., the Linux kernel, Mozilla, and Apache, for three important types of bugs: buffer overflows, null pointer bugs, and memory leaks. R2Fix generates 60 patches correctly, 5 of which are new patches for bugs that have not been fixed by developers yet. We reported all 5 new patches to the developers; 4 have already been accepted and committed to the code repositories. The 60 correct patches generated by R2Fix could have shortened and saved an average of 68 days of bug diagnosis and patch generation time.

Many software companies use issue tracking systems to organize their work. However, when working on large projects, across multiple teams, a problem of finding the correctteam to solve a certain issue arises. One team might detect a problem, which must be solved by another team. This can take time from employees tasked with finding the correct team and automating the dispatching of these issues can have large benefits for the company. In this thesis, the use of machine learning methods, mainly convolutional neural networks (CNN) for text classification, has been applied to this problem. For natural language processing both word- and character-level representations are commonly used. The results in this thesis suggests that the CNN learns different information based on whether word- or character-level representation is used. Furthermore, it was concluded that the CNN models performed on similar levels as the classical Support Vector Machine for this task. When compared to a human expert, working with dispatching issues, the best CNN model performed on a similar level when given the same information. The high throughput of a computer model, therefore, suggests automation of this task is very much possible.

Today’s world is a network of interconnected systems that are always running to facilitate information exchange so people can carry out their daily activities. Software applications are constantly evolving to meet the increasing expectations of the growing market, thereby giving rise to the development of large complex systems. It is very likely for these complex systems to encounter bugs which is a situation that can cause errors in software. These bugs can prevent the systems from operating as intended, slowing down software development and deployment, and causing delays in deadlines. This study undertook a systematic literature review to find trends in the field of bug prioritization. Software bug prioritization can help developers determine the order of fixing bugs by assigning priority levels based on the severity analysis. This study aims to identify the most promising techniques that can change the bug prediction and resolution process. It is observed that machine learning techniques (ML) have been gaining popularity in addressing the bug prioritization issue since they can automatically assign priority levels. However, these ML techniques also have limitations addressed in this study along with a taxonomic classification of identified techniques. The review obtained 34 manuscripts based on study selection criteria. These manuscripts discovered 63 unique bug prioritization techniques, including a mix of ML, data reduction and hybrid techniques. It is evident that though these techniques perform automatic prioritization, they can sometimes be slow and lack consistency in the accuracy of results.

For the task of near-duplicate document detection, comparison approaches based on bag-of-words used in information retrieval community are not sufficiently accurate. This work presents novel approach when instance-level constraints are given for documents and it is needed to retrieve them, given new query document for near-duplicate detection. The framework incorporates instance-level constraints and clusters documents into groups using novel clustering approach Grouped Latent Dirichlet Allocation (gLDA). Then distance metric is learned for each cluster using large margin nearest neighbor algorithm and finally ranked documents for given new unknown document using learnt distance metrics. The variety of experimental results on various datasets demonstrate that our clustering method (gLDA with side constraints) performs better than other clustering methods and the overall approach outperforms other near-duplicate detection algorithms.