No description available.

Computing education for learning to program has made great strides in the current century. Exciting educational technologies are now available and active learning pedagogies are increasingly used. Interest is strong, but the longstanding problem remains: learning to program as an analytical endeavor is quite frustrating for many. The purpose of this study is to discover ways to mitigate this frustration. It researches ways to help students comprehend code by guiding them to take it apart (through reading, tracing, completing, and debugging) as they learn to write code on their own. This study contributes to the understanding of learning from errors. It also builds upon and further develops the emergent pedagogy of de-constructionism. The de-constructionist approach involves taking things apart, practice, and learning from errors. This study applies a de-constructionist approach in an experiment with ~80 undergraduates learning Python in an introductory programming class. During weekly lab periods, students engaged with web-based interactive practice problems that emphasize reading, tracing, completing, and in some cases, debugging code. Students also wrote code for lab and homework assignments. Approximately half of the students were given some that involved learning from bugs that were intentionally placed in the provided code, while the others were not. Learning gains were assessed using pre/post tests and exams. Surveys were used to measure attitudes. Learning gains and attitudes were compared according to condition (Bugs, NoBugs), prior experience, gender, minority status, and class size. This study demonstrates that bugs can be intentionally incorporated into practice problems that students like to solve, without detrimental effects on learning or attitudes about computing. It also contributes to the literature on code comprehension. / Math & Science Education

The purpose of this study is to analyze which behaviors are or are not helpful for debugging when a novice is in a state of unproductive persistence. Further, this project will exploratorily use a variety of analytical techniques -- including association rule mining, process mining, frequent sequence mining, and machine learning-- in order to determine which approaches are useful for data analysis. For the study, programming process data from hundreds of novice programmers were analyzed to determine which behaviors were more or less likely to be correlated with escaping a state of unproductive persistence. Of these events, only three had a statistically significant difference in their rates of occurrence and large effect sizes: file, edit, and compile events. While the data set cannot reveal a user's motivation for a file event, the most logical explanation of these events is that the user is tracing the code. Thus, a higher rate of file events suggests that code tracing (with the goal of code comprehension) is a key behavior correlated with a student's ability to escape a state of unproductive persistence. On the other hand, editing events are far more common in unproductive states that are not escaped. A content analysis suggests that there are more trivial edits for users in an unescaped state of unproductive persistence. An important finding of this study is that an unproductive persistence is not just a phenomenon of the worst-performing students; rather, a third of users who completed the assignment had at least one unproductive state. This study also lends support to the idea that tinkering combined with code tracing is correlated with positive outcomes, but that less systematic tinkering is not effective behavior. Further, association rule mining and frequent sequence mining were effective tools for data analysis in this study. The findings from this study have two main practical implications for curriculum designers and instructors: (1) the need to normalize struggle and (2) possibilities for curriculum and tool development. This work is particularly important given that debugging is not normally a process evident to instructors, curriculum designers, tool developers, and computer science education researchers, either because it happens outside of class time and/or because it is a process and these stakeholders usually only see the end result; this project attempts to make the process of debugging more transparent.

Our reliance on software has been growing fast over the past decades as the pervasive use of computer and software penetrated not only our daily life but also many critical applications. As the computational power of multi-core processors and other parallel hardware keeps increasing, concurrent software that exploit these parallel computing hardware become crucial for achieving high performance. However, developing correct and efficient concurrent software is a difficult task for programmers due to the inherent nondeterminism in their executions. As a result, concurrency related software bugs are among the most troublesome in practice and have caused severe problems in recent years. In this dissertation, I propose a series of new and fully automated methods for verifying and debugging concurrent software. They cover the detection, prevention, classification, and repair of some important types of bugs in the implementation of concurrent data structures and client-side web applications. These methods can be adopted at various stages of the software development life cycle, to help programmers write concurrent software correctly as well as efficiently. / Ph. D.

Debugging is a challenging and time-consuming process in software life-cycle. The focus of the thesis is to improve the accuracy of existing fault localization (FL) techniques. We experimented with several source code line level features such as line commit size, line recency, and line length to arrive at a new fault localization technique. Based on our experiments, we propose a novel enhanced cost-aware fault localization (ECFL) technique by combining line length with the existing selected baseline fault localization techniques. ECFL improves the accuracy of DStar (Baseline 1), CombineFastestFL (Baseline 2), and CombineFL (Baseline 3) by locating 81%, 58%, and 30% more real faults respectively in Top-1 evaluation metric. In comparison with the baseline techniques, ECFL requires a marginal additional time (on an average, 5 seconds per bug) and data while providing a significant improvement in accuracy. The source code line features also improve the baseline fault localization techniques when ''learning to rank'' SVM machine learning approach is used to combine the features. We also provide an infrastructure to facilitate future research on combining new source code line features with other fault localization techniques. / Master of Science / Software debugging involves locating and fixing faults (or bugs) in software. It is a challenging and time-consuming process in software life-cycle. Fault localization (FL) techniques help software developers to locate faults by providing a ranked set of program elements. The focus of the thesis is to improve the accuracy of existing fault localization techniques. We experimented with several source code line level features such as line commit size, line recency, and line length to arrive at a new fault localization technique. Based on our experiments, we propose a novel enhanced cost-aware fault localization (ECFL) technique by combining line length with the existing selected baseline fault localization techniques. ECFL improves the accuracy of DStar (Baseline 1), CombineFastestFL (Baseline 2), and CombineFL (Baseline 3) by locating 81%, 58%, and 30% more real faults respectively in Top-1 evaluation metric. In comparison with the baseline techniques, ECFL requires a marginal additional time (on an average, 5 seconds per bug) and data while providing a significant improvement in accuracy. The source code line features also improve the baseline fault localization techniques when machine learning approach is used to combine the features. We also provide an infrastructure to facilitate future research on combining new source code line features with other fault localization techniques.

Automated grading systems provide feedback to computer science students in a variety of ways, but often focus on incorrect program behaviors. These tools will provide indications of test case failures or runtime errors, but without debugging skills, students often become frus- trated when they don't know where to start. They know their code has defects, but finding the problem may be beyond their experience, especially for beginners. An additional concern is balancing the need to provide enough direction to be useful, without giving the student so much direction that they are effectively given the answer. Drawing on the experiences of the software engineering community, in this work we apply a technique called statistical fault location (SFL) to student program assignments. Using the GZoltar software tool, we applied this technique to a set of previously-submitted student assignments gathered from students in our introductory CS course, CS 1114: Introduction to Software Design. After a manual inspection of the student code, this exercise demonstrated that the SFL technique identifies the defective method in the first three most suspicious methods in the student's code 90% of the time. We then developed a plug-in for Web-CAT to allow new student submissions to be evaluated with the GZoltar SFL system. Additionally, we developed a tool to create a heat map visualization to show the results of the SFL evaluation overlaid on the student's source code. We deployed this toolset for use in CS 1114 in Fall 2017. We then surveyed the students about their perceptions of the utility of the visualization for helping them understand how to find and correct the defects in their code, versus not having access to the heat map. Their responses led to refinements in our presentation of the feedback. We also evaluated the performance of CS 1114 classes from two semesters and discovered that having the heat maps led to more frequent incremental improvements in their code, as well as reaching their highest correctness score on instructor-provided tests more quickly than students that did not have access to the heat maps. Finally, we suggest several directions for future enhancements to the feedback interface. / Doctor of Philosophy / Automated grading systems provide feedback to computer science students in a variety of ways, but often focus on incorrect program behaviors. These tools will provide indications of test case failures or runtime errors, but without debugging skills, students often become frus- trated when they don't know where to start. They know their code has defects, but finding the problem may be beyond their experience, especially for beginners. An additional concern is balancing the need to provide enough direction to be useful, without giving the student so much direction that they are effectively given the answer. Drawing on the experiences of the software engineering community, in this work we apply a technique called statistical fault location (SFL) to student program assignments. Using the GZoltar software tool, we applied this technique to a set of previously-submitted student assignments gathered from students in our introductory CS course, CS 1114: Introduction to Software Design. After a manual inspection of the student code, this exercise demonstrated that the SFL technique identifies the defective method in the first three most suspicious methods in the student's code 90% of the time. We then developed a plug-in for Web-CAT to allow new student submissions to be evaluated with the GZoltar SFL system. Additionally, we developed a tool to create a heat map visualization to show the results of the SFL evaluation overlaid on the student's source code. We deployed this toolset for use in CS 1114 in Fall 2017. We then surveyed the students about their perceptions of the utility of the visualization for helping them understand how to find and correct the defects in their code, versus not having access to the heat map. Their responses led to refinements in our presentation of the feedback. We also evaluated the performance of CS 1114 classes from two semesters and discovered that having the heat maps led to more frequent incremental improvements in their code, as well as reaching their highest correctness score on instructor-provided tests more quickly than students that did not have access to the heat maps. Finally, we suggest several directions for future enhancements to the feedback interface.

The increasing complexity of Web system has led to the development of sophisticated formal methodologies for verifying and correcting Web data and Web programs. In general, establishing whether a Web system behaves correctly with respect to the original intention of the programmer or checking its internal consistency are non-trivial tasks as witnessed by many studies in the literature. In this dissertation, we face two challenging problems related to the verification of Web systems. Firstly, we extend a previous Web verification framework based on partial rewriting by providing a semi-automatic technique for repairing Web systems. We propose a basic repairing methodology that is endowed with several strategies for optimizing the number of repair actions that must be executed in order to fix a given Web site. Also, we develop an improvement of the Web verification framework that is based on abstract interpretation and greatly enhances both efficiency and scalability of the original technique. Secondly, we formalize a framework for the specification and model-checking of dynamic Web applications that is based on Rewriting Logic. Our framework allows one to simulate the user navigation and the evaluation of Web scripts within a Web application, and also check important related properties such us reachability and consistency. When a property is refuted, a counter-example with the erroneous trace is delivered. Such information can be analyzed in order to debug the Web application under examination by means of a novel backward trace slicing technique that we formulated for this purpose. This technique consists in tracing back, along an execution trace, all the relevant symbols of the term (or state) that we are interested to observe. / Romero ., DO. (2011). Rewriting-based Verification and Debugging of Web Systems [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/12496

In a society that is increasingly dependent upon computing machinery, the issues associated with the correct functioning of that machinery are of crucial interest. The consequences of erroneous behavior of computers are dire with the worst case scenario being, conceivably, global thermonuclear war. Therefore, development of procedures and tools which can be used to increase the confidence of the correctness of the software that controls the world's computers is of vital importance. The Department of Defense (DoD) is in the process of adopting a standard computer language for the development of software. This language is called Ada¹. One of the major features of Ada is that it supports concurrent programming via its "task" compilation unit. There are not, however, any automated tools to aid in locating errors in the tasks. The design for such a tool is presented. The tool is named AdaTAD and is a debugger for programs written in Ada. The features of AdaTAD are specific to the problems of concurrent programming. The requirements of AdaTAD are derived from the literature. AdaTAD is, however, a unique tool designed using Ada as a program description language. When AdaTAD is implemented in Ada it becomes portable among all environments which support the Ada language. This offers the advantage that a single debugger is portable to many different machine architectures. Therefore, separate debuggers are not necessary for each implementation of Ada. Moreover, since AdaTAD is designed to allow debugging of tasks, AdaTAD will also support debugging in a distributed environment. That means that, if the tasks of a user's program are running on different computers in a distributed environment, the user is still able to use AdaTAD to debug the tasks as a single program. This feature is unique among automated debuggers. After the design is presented, several examples are offered to explain the operation of AdaTAD and to show that AdaTAD is useful in revealing the location of errors specific to concurrent programming. / Ph. D.

<p dir="ltr">Tree-based machine learning models, such as decision trees and random forests, are one of the most widely used machine learning models primarily because of their predictive power in supervised learning tasks and ease of interpretation. Despite their popularity and power, these models have been found to produce unexpected or discriminatory behavior. Given their overwhelming success for most tasks, it is of interest to identify root causes of the unexpected and discriminatory behavior of tree-based models. However, there has not been much work on understanding and debugging tree-based classifiers in the context of fairness. We introduce FairDebugger, a system that utilizes recent advances in machine unlearning research to determine training data subsets responsible for model unfairness. Given a tree-based model learned on a training dataset, FairDebugger identifies the top-k training data subsets responsible for model unfairness, or bias, by measuring the change in model parameters when parts of the underlying training data are removed. We describe the architecture of FairDebugger and walk through real-world use cases to demonstrate how FairDebugger detects these patterns and their explanations.</p>

The aim of this work is to integrate Java PathFinder into Java Platform Debugger Architecture. That is, to allow using Java PathFinder instead of a common Java Virtual Machine for the purpose of debugging Java applications in any modern Java Integrated Development Environment with all its advantages such as various kinds of breakpoints, direct stepping in opened source files, and call stack and object inspection. The resulting work provides users with all the features they are used to while debugging Java applications. None of this requires any external tools, editors or a complicated setup. Therefore, users are able to view, debug and understand the program state while replaying an error trace in Java PathFinder. The key part of the study is an implementation of the Java Debug Wire Protocol Agent as an extension for Java PathFinder. That makes JPF more complete as a Virtual Machine in the eyes of the community and the Java users in general. Powered by TCPDF (www.tcpdf.org)