Search

Search Results

• 1. [Article] How well do professional developers test with code coverage visualizations? An empirical study

  "Code coverage visualizations using block coverage neither guided developers toward productive testing strategies, nor did these visualizations motivate developers to write more tests or help them find ...

  Citation × Citation Citation Abstract Copy Title: How well do professional developers test with code coverage visualizations? An empirical study Author: Burnett, Margaret, 1949-, Lawrance, Joseph A., Clarke, Steven, Oregon State University. Dept. of Computer Science, Rothermel, Gregg "Code coverage visualizations using block coverage neither guided developers toward productive testing strategies, nor did these visualizations motivate developers to write more tests or help them find more faults than the control group. Nevertheless, code coverage visualizations did influence developers in a few important ways. Code coverage visualizations led developers to overestimate their test effectiveness more than the control group. Yet, these same visualizations reduced the variability in the number of test cases developers wrote by changing the standard developers used to evaluate their test effectiveness. Thus, the true power of testing visualizations lies not only with the faults that visualizations can highlight; it also lies in how visualizations can change how developers think about testing. Testing visualizations guide developers toward a particular standard of effectiveness, so if we want developers to test software adequately, we must ensure that the coverage criteria we choose to visualize leads developers toward a good standard of test effectiveness."--Conclusion Title: How well do professional developers test with code coverage visualizations? An empirical study Author: Burnett, Margaret, 1949-, Lawrance, Joseph A., Clarke, Steven, Oregon State University. Dept. of Computer Science, Rothermel, Gregg "Code coverage visualizations using block coverage neither guided developers toward productive testing strategies, nor did these visualizations motivate developers to write more tests or help them find more faults than the control group. Nevertheless, code coverage visualizations did influence developers in a few important ways. Code coverage visualizations led developers to overestimate their test effectiveness more than the control group. Yet, these same visualizations reduced the variability in the number of test cases developers wrote by changing the standard developers used to evaluate their test effectiveness. Thus, the true power of testing visualizations lies not only with the faults that visualizations can highlight; it also lies in how visualizations can change how developers think about testing. Testing visualizations guide developers toward a particular standard of effectiveness, so if we want developers to test software adequately, we must ensure that the coverage criteria we choose to visualize leads developers toward a good standard of test effectiveness."--Conclusion Close

• 2. [Article] Continuous Integration (CI) Needs and Wishes for Developers of Proprietary Code

  Continuous integration (CI) systems automate the compilation, building, and testing of software. Despite CI being one of the most widely used processes in software engineering, we do not know what motivates ...

  Citation × Citation Citation Abstract Copy Title: Continuous Integration (CI) Needs and Wishes for Developers of Proprietary Code Author: Dig, Danny, Tunnell, Timothy, Nelson, Nicholas, Oregon State University. Dept. of Computer Science, Hilton, Michael, Marinov, Darko Continuous integration (CI) systems automate the compilation, building, and testing of software. Despite CI being one of the most widely used processes in software engineering, we do not know what motivates developers to use CI, and what barriers and unmet needs they face. Without such knowledge developers make easily avoidable errors, managers reduce the productivity of developers by making misinformed decisions, tool builders invest in the wrong direction, and researchers miss many opportunities for improving the software engineering practice. Given the large fraction of proprietary code development, understanding how proprietary developers are using CI is vital to improving it. We present the first study of how CI is used in the proprietary development of software. We conduct 16 semi-structured interviews with developers from different industries and development scale. We generalize these findings by surveying 523 developers from all over the world. We find that 78% of developers feel more productive when using CI, and 85% think that using CI causes developers to give more value to automated testing. However, unlike open-source developers, many proprietary developers want to use CI but are not allowed to do so. 50% of developers have problems troubleshooting CI builds, and more than half of developers want easier configurations for CI tools and services. Title: Continuous Integration (CI) Needs and Wishes for Developers of Proprietary Code Author: Dig, Danny, Tunnell, Timothy, Nelson, Nicholas, Oregon State University. Dept. of Computer Science, Hilton, Michael, Marinov, Darko Continuous integration (CI) systems automate the compilation, building, and testing of software. Despite CI being one of the most widely used processes in software engineering, we do not know what motivates developers to use CI, and what barriers and unmet needs they face. Without such knowledge developers make easily avoidable errors, managers reduce the productivity of developers by making misinformed decisions, tool builders invest in the wrong direction, and researchers miss many opportunities for improving the software engineering practice. Given the large fraction of proprietary code development, understanding how proprietary developers are using CI is vital to improving it. We present the first study of how CI is used in the proprietary development of software. We conduct 16 semi-structured interviews with developers from different industries and development scale. We generalize these findings by surveying 523 developers from all over the world. We find that 78% of developers feel more productive when using CI, and 85% think that using CI causes developers to give more value to automated testing. However, unlike open-source developers, many proprietary developers want to use CI but are not allowed to do so. 50% of developers have problems troubleshooting CI builds, and more than half of developers want easier configurations for CI tools and services. Close

• 3. [Article] A meta-model for software process maturity

  We present a meta-model as a process maturity framework that can be effectively used to identify the key practices to initiate and sustain a software process improvement effort focused on a single process ...

  Citation × Citation Citation Abstract Copy Title: A meta-model for software process maturity Author: Cook, Curtis R., Visconti, Marcello, Oregon State University. Dept. of Computer Science We present a meta-model as a process maturity framework that can be effectively used to identify the key practices to initiate and sustain a software process improvement effort focused on a single process area. Our approach moves away from the overall software development process and computation of maturity levels and focuses on a particular process area or task and its key practices. In determining process maturity our framework takes into account not only the process dimension but also the quality and usability of the products or deliverables of the process. We demonstrate the applicability of the proposed meta-model by generating key practices for several software process areas and tasks including system documentation, quality assurance and configuration management. We show that our approach is able to identify a complete set of key practices to assess a process and generate recommendations that facilitate the development of an action plan for the software process improvement effort. Title: A meta-model for software process maturity Author: Cook, Curtis R., Visconti, Marcello, Oregon State University. Dept. of Computer Science We present a meta-model as a process maturity framework that can be effectively used to identify the key practices to initiate and sustain a software process improvement effort focused on a single process area. Our approach moves away from the overall software development process and computation of maturity levels and focuses on a particular process area or task and its key practices. In determining process maturity our framework takes into account not only the process dimension but also the quality and usability of the products or deliverables of the process. We demonstrate the applicability of the proposed meta-model by generating key practices for several software process areas and tasks including system documentation, quality assurance and configuration management. We show that our approach is able to identify a complete set of key practices to assess a process and generate recommendations that facilitate the development of an action plan for the software process improvement effort. Close

• 4. [Article] On to the real world : gender and self-efficacy in Excel

  Although there have been a number of studies of end-user software development tasks, few of them have considered gender issues for real end-user developers in real-world environments for end-user programming. In ...

  Citation × Citation Citation Abstract Copy Title: On to the real world : gender and self-efficacy in Excel Author: Inman, Derek, Rector, Kyle, Beckwith, Laura A., Burnett, Margaret, 1949-, Oregon State University. Dept. of Computer Science Although there have been a number of studies of end-user software development tasks, few of them have considered gender issues for real end-user developers in real-world environments for end-user programming. In order to be trusted, the results of such laboratory studies must always be re-evaluated with fewer controls, more closely reflecting real-world conditions. Therefore, the research question in this paper is whether the results of a Gender HCI controlled study generalize -- to real-world end-user developers, in a real-world spreadsheet environment, using a realworld spreadsheet. Our findings are that the concepts revealed by the original laboratory study appear to be quite robust, being demonstrated in multiple ways in this real-world environment. Title: On to the real world : gender and self-efficacy in Excel Author: Inman, Derek, Rector, Kyle, Beckwith, Laura A., Burnett, Margaret, 1949-, Oregon State University. Dept. of Computer Science Although there have been a number of studies of end-user software development tasks, few of them have considered gender issues for real end-user developers in real-world environments for end-user programming. In order to be trusted, the results of such laboratory studies must always be re-evaluated with fewer controls, more closely reflecting real-world conditions. Therefore, the research question in this paper is whether the results of a Gender HCI controlled study generalize -- to real-world end-user developers, in a real-world spreadsheet environment, using a realworld spreadsheet. Our findings are that the concepts revealed by the original laboratory study appear to be quite robust, being demonstrated in multiple ways in this real-world environment. Close

• 5. [Article] Gencel : a program generator for correct spreadsheets

  A huge discrepancy between theory and practice exists in one popular application area of functional programming--spreadsheets. Although spreadsheets are the most frequently used (functional) programs, ...

  Citation × Citation Citation Abstract Copy Title: Gencel : a program generator for correct spreadsheets Author: Cooperstein, Irene, Kollmansberger, Steven, Erwig, Martin, Oregon State University. Dept. of Computer Science, Abraham, Robin A huge discrepancy between theory and practice exists in one popular application area of functional programming--spreadsheets. Although spreadsheets are the most frequently used (functional) programs, few formal models of computation and type systems exist that would provide the foundation for creating reliable spreadsheets. Consequently, existing spreadsheets contain many errors, some of which have serious impacts. We argue in favor of creating a formal foundation for spreadsheets to help improve spreadsheet systems and make spreadsheets more reliable software assets. To this end, we have developed a table specification language that allows the definition of table templates that describe possible spreadsheet evolutions. This language is based on a table calculus that formally captures the process of creating and modifying spreadsheets. We have developed a type system for this calculus that can prevent type, reference, and omission errors from occurring in spreadsheets. On the basis of the table calculus we have developed Gencel, a system for generating reliable spreadsheets. We have implemented a prototype version of Gencel as an extension of Excel. A pilot study has indicated that the system is well accepted by end-user programmers. Title: Gencel : a program generator for correct spreadsheets Author: Cooperstein, Irene, Kollmansberger, Steven, Erwig, Martin, Oregon State University. Dept. of Computer Science, Abraham, Robin A huge discrepancy between theory and practice exists in one popular application area of functional programming--spreadsheets. Although spreadsheets are the most frequently used (functional) programs, few formal models of computation and type systems exist that would provide the foundation for creating reliable spreadsheets. Consequently, existing spreadsheets contain many errors, some of which have serious impacts. We argue in favor of creating a formal foundation for spreadsheets to help improve spreadsheet systems and make spreadsheets more reliable software assets. To this end, we have developed a table specification language that allows the definition of table templates that describe possible spreadsheet evolutions. This language is based on a table calculus that formally captures the process of creating and modifying spreadsheets. We have developed a type system for this calculus that can prevent type, reference, and omission errors from occurring in spreadsheets. On the basis of the table calculus we have developed Gencel, a system for generating reliable spreadsheets. We have implemented a prototype version of Gencel as an extension of Excel. A pilot study has indicated that the system is well accepted by end-user programmers. Close

• 6. [Article] Workshop report : from end-user programming to end-user software engineering (a CHI'06) Workshop)

  End users create software when they use spreadsheet systems, web authoring tools and graphical languages, and when they create educational simulations, macros-by-demonstration, and dynamic e-business web ...

  Citation × Citation Citation Abstract Copy Title: Workshop report : from end-user programming to end-user software engineering (a CHI'06) Workshop) Author: Myers, Brad A., Burnett, Margaret, 1949-, Rosson, Mary Beth, Oregon State University. Dept. of Computer Science, Wiedenbeck, Susan, Leibel, Adam End users create software when they use spreadsheet systems, web authoring tools and graphical languages, and when they create educational simulations, macros-by-demonstration, and dynamic e-business web applications and mash-ups. Some end-user developers, such as accountants or teachers, may have no formal training at all in programming. Others, such as scientists or sysadmins, may have significant background in programming, but they still do not see their goals as producing software: rather, they see software as a means to some other goal, such as visualizing a scientific phenomenon or getting a computer set up so that it can talk to the printers in the building. It has become well known that errors are pervasive in end-user software, and the resulting impact is sometimes enormous. A growing number of researchers and developers are working on ways to make the software created by end users more reliable, by looking beyond the "create" aspect of the software lifecycle. The WEUSE workshop (Workshop on End-User Software Engineering), organized by Margaret Burnett, Brad Myers, Mary Beth Rosson, and Susan Wiedenbeck for CHI'06, brought together academic and industrial researchers who are addressing this topic with industry representatives who are deploying end-user programming systems. The objective was to facilitate sharing of real-world problems and solutions. Four real-world end-user software engineering issues formed the focus of the CHI'06 workshop: extending an open source paradigm for end-user developers of MATLAB programs, helping disparate audiences learn to use the Flash end-user programming product, dealing with mission-critical aspects of end-user programming situations that occur in customization of healthcare systems, and supporting the high-pressure development done by sysadmins. Title: Workshop report : from end-user programming to end-user software engineering (a CHI'06) Workshop) Author: Myers, Brad A., Burnett, Margaret, 1949-, Rosson, Mary Beth, Oregon State University. Dept. of Computer Science, Wiedenbeck, Susan, Leibel, Adam End users create software when they use spreadsheet systems, web authoring tools and graphical languages, and when they create educational simulations, macros-by-demonstration, and dynamic e-business web applications and mash-ups. Some end-user developers, such as accountants or teachers, may have no formal training at all in programming. Others, such as scientists or sysadmins, may have significant background in programming, but they still do not see their goals as producing software: rather, they see software as a means to some other goal, such as visualizing a scientific phenomenon or getting a computer set up so that it can talk to the printers in the building. It has become well known that errors are pervasive in end-user software, and the resulting impact is sometimes enormous. A growing number of researchers and developers are working on ways to make the software created by end users more reliable, by looking beyond the "create" aspect of the software lifecycle. The WEUSE workshop (Workshop on End-User Software Engineering), organized by Margaret Burnett, Brad Myers, Mary Beth Rosson, and Susan Wiedenbeck for CHI'06, brought together academic and industrial researchers who are addressing this topic with industry representatives who are deploying end-user programming systems. The objective was to facilitate sharing of real-world problems and solutions. Four real-world end-user software engineering issues formed the focus of the CHI'06 workshop: extending an open source paradigm for end-user developers of MATLAB programs, helping disparate audiences learn to use the Flash end-user programming product, dealing with mission-critical aspects of end-user programming situations that occur in customization of healthcare systems, and supporting the high-pressure development done by sysadmins. Close

• 7. [Article] Multiparadigm data structures in Leda

  Multiparadigm programming is a term used to describe a style of software development that makes use of facilities originally designed in support of a number of different programming language paradigms. ...

  Citation × Citation Citation Abstract Copy Title: Multiparadigm data structures in Leda Author: Budd, Timothy, Oregon State University. Dept. of Computer Science Multiparadigm programming is a term used to describe a style of software development that makes use of facilities originally designed in support of a number of different programming language paradigms. In this paper we illustrate our conception of multiparadigm programming, by describing how various data structures can be implemented in the programming language Leda. Leda is a strongly-typed compiled multiparadigm programming language that we have been developing over the past several years. Our exposition serves both to illustrate the idea of multiparadigm programming, and to describe the features of the language Leda. Title: Multiparadigm data structures in Leda Author: Budd, Timothy, Oregon State University. Dept. of Computer Science Multiparadigm programming is a term used to describe a style of software development that makes use of facilities originally designed in support of a number of different programming language paradigms. In this paper we illustrate our conception of multiparadigm programming, by describing how various data structures can be implemented in the programming language Leda. Leda is a strongly-typed compiled multiparadigm programming language that we have been developing over the past several years. Our exposition serves both to illustrate the idea of multiparadigm programming, and to describe the features of the language Leda. Close

• 8. [Article] An integrated testing and fault localization methodology for spreadsheet languages

  Spreadsheet languages, which include commercial spreadsheets and various research systems, have proven to be flexible tools in many settings. Research shows, however, that spreadsheets often contain faults. ...

  Citation × Citation Citation Abstract Copy Title: An integrated testing and fault localization methodology for spreadsheet languages Author: Burnett, Margaret, 1949-, Reichwein, James, Oregon State University. Dept. of Computer Science Spreadsheet languages, which include commercial spreadsheets and various research systems, have proven to be flexible tools in many settings. Research shows, however, that spreadsheets often contain faults. This thesis presents an integrated testing and fault localization methodology for spreadsheets. This methodology allows spreadsheet developers to engage in modeless development, testing and debugging activities. Furthermore, we provide an interface to our methodology that does not require an understanding of testing and debugging theory. To accomplish this, we introduce the notion of fault likelihood: the likelihood that a given cell contains a fault that contributes to an known failure in the spreadsheet. To estimate fault likelihood we present five properties that we feel should govern its behavior. We then discuss our implementation of this methodology and illustrate its use. Title: An integrated testing and fault localization methodology for spreadsheet languages Author: Burnett, Margaret, 1949-, Reichwein, James, Oregon State University. Dept. of Computer Science Spreadsheet languages, which include commercial spreadsheets and various research systems, have proven to be flexible tools in many settings. Research shows, however, that spreadsheets often contain faults. This thesis presents an integrated testing and fault localization methodology for spreadsheets. This methodology allows spreadsheet developers to engage in modeless development, testing and debugging activities. Furthermore, we provide an interface to our methodology that does not require an understanding of testing and debugging theory. To accomplish this, we introduce the notion of fault likelihood: the likelihood that a given cell contains a fault that contributes to an known failure in the spreadsheet. To estimate fault likelihood we present five properties that we feel should govern its behavior. We then discuss our implementation of this methodology and illustrate its use. Close

• 9. [Article] Approximation algorithms for solving cost observable Markov decision processes

  "The specifi c problem addressed in this proposal is the development of good approximation algorithms for solving problems that have partial observability. The model we propose associates costs with obtaining ...

  Citation × Citation Citation Abstract Copy Title: Approximation algorithms for solving cost observable Markov decision processes Author: Bayer, Valentina, Oregon State University. Dept. of Computer Science "The specifi c problem addressed in this proposal is the development of good approximation algorithms for solving problems that have partial observability. The model we propose associates costs with obtaining information about the current state. We want to predict when and how much it is necessary to observe. We want to use our Cost Observable Markov Decision Process (COMDP) model to find good solutions for real-world problems ..."--Problem definition. Title: Approximation algorithms for solving cost observable Markov decision processes Author: Bayer, Valentina, Oregon State University. Dept. of Computer Science "The specifi c problem addressed in this proposal is the development of good approximation algorithms for solving problems that have partial observability. The model we propose associates costs with obtaining information about the current state. We want to predict when and how much it is necessary to observe. We want to use our Cost Observable Markov Decision Process (COMDP) model to find good solutions for real-world problems ..."--Problem definition. Close

• 10. [Article] A new approach to vector code generation for applicative languages

  In an earlier paper we developed an intermediate representation for languages based on composition, and showed how the representation could facilitate generating code for functional languages, such as ...

  Citation × Citation Citation Abstract Copy Title: A new approach to vector code generation for applicative languages Author: Budd, Timothy, Oregon State University. Dept. of Computer Science In an earlier paper we developed an intermediate representation for languages based on composition, and showed how the representation could facilitate generating code for functional languages, such as FP. In this paper we follow the same philosophical approach, using instead the applicative language APL. Further, we show how this intermediate representation simplifies the task of generating code for highly parallel machines, such as the Connection Machine. Title: A new approach to vector code generation for applicative languages Author: Budd, Timothy, Oregon State University. Dept. of Computer Science In an earlier paper we developed an intermediate representation for languages based on composition, and showed how the representation could facilitate generating code for functional languages, such as FP. In this paper we follow the same philosophical approach, using instead the applicative language APL. Further, we show how this intermediate representation simplifies the task of generating code for highly parallel machines, such as the Connection Machine. Close