![[PukiWiki]](image/wadanyan.png "[PukiWiki]")
#author("2019-01-23T10:30:22+09:00","","") #author("2019-01-23T10:31:49+09:00","","") [[論文英語単語帳/表現集]] ***リンク [#f20ebad6] #ls2(論文英語単語帳/) ** 表現集[#s089e92b] #ls2(論文英語単語帳/表現集/) *近いものとして [#ec5d16b9] -closest to --Closest to our work, is the series of work on bug report severity prediction by Menzies and Marcus [16], Lamkanfi et al. [13], [14], and our own previous work [27]. -closer to --Closer to our work, in as much as some prediction is being made based on observed rules, are the approaches that use the textual similarity between bug descriptions to assign bugs to developers -Closer XXX --Closer analysis showed that Perl uses a specific semantic versioning approach [27] where even release numbers, like 5.8 and 5.10, are official production releases (with maintenance releases, such as 5.12.1 and 5.12.2 mainly for bug fixes) and odd release numbers, like 5.11 and 5.13, are development releases. *がある [#tb4a3d6a] -There exists --There exists previous work related to build, build failures, and CPAN or other ecosystems. -conducted --Empirical studies of bug characteristics have been conducted in the past to learn about concurrency bugs [27, 33], performance bugs [23, 40], and error handling bugs [13, 16], among many others. *Aは〜した [#kfbf9117] **調査 [#i72773b4] -investigate --Denny et al. [7] investigated compile errors in short pieces of Java code and how students fixed these errors. They showed that 48% of the builds failed due to compilation errors. -study --Suvorov et al. [9] studied two popular open-source projects, Linux and KDE, and found that the migration from one build system to another might fail due to missing features. --The evolution and maintenance of build system specification files primarily has been studied quantitatively [33, 48, 78]. -surveyed --Stahle and Bosch [13] surveyed CI practices and build failures in CI systems. -conduct a study --Vasilescu et al. [15] conducted an empirical study about CI usage on 246 projects from GitHub. -attempt --However, no studies have attempted to study the same phenomena in fault prediction models. **評価 [#x8830104] -assess --Similarly, Dyke [8] assessed the frequency of compile errors by tracking Eclipse IDE usage with novice programmers. --Hilton et al. [16] assessed 34,544 open-source projects from GitHub, 40% of which use CI. **分析 [#n36019a0] -analyze --Adams et al. [3], [10] analyzed the changes to the Linux kernel build system and reported that the build system grew and had to frequently be maintained. **リプリケートスタディ [#pa8539ca] -replicate --McIntosh et al. [11] replicated the same study on the ANT build system. We complement these studies with our predictive model of build failures for the build system of CPAN. **フォローアップ [#w6b564d1] -follow up --Wu et al. [23] followed up on this first study by analyzing changes in 22 releases of the Apache and Eclipse frameworks and their client programs. **報告 [#s47cacc4] -report --Seo et al. [4] reported on a case study performed at Google, in which they assessed 26.6 million builds in C and Java. *分かった，示した [#w077e497] -found --Suvorov et al. [9] studied two popular open-source projects, Linux and KDE, and found that the migration from one build system to another might fail due to missing features. -observe --They observed that test failures during builds are sometimes accepted by developers because developers known that these particular failures will be fixed later [14]. --They observed the kind of API changes in the frameworks impacting the clients and classified API changes and API usages. -show -- They showed that there exist multiple patterns of build failures and focused on compilation problems to conclude that dependencies are the main source of compilation errors, that the time to fix these errors vary widely, and that developers need more, different tools support. -suggest --The recent study of Gil and Lalouche [9] suggested that code size is the only unique valid metric and therefore fault prediction models should not in fact control for size. -explain --The authors explained that the more a metric is correlated with size, the more able it is to predict bugs. -emphasize --The need to adopt more robust statistical methods in empirical software engineering research has been emphasized recently by Kitchenham et al. [17], as has the (consequent) need to reassess evidence from prior empirical results. **既存研究との違いを述べる [#w052530d] -complement --We complement this study that is showing the impact of OSes and environments on build failures within CPAN particular unnamed CI system.” -provide evidence supporting --We provide evidence supporting Seo et al. [4] on the inflation of the number of builds. **既存研究の問題を述べる [#sf950713] -there are problems with In particular, there are problems with the assumption of normality of data distributions, and also regarding the power and Type I error control of the tests employed in both studies. As such their conclusions may not hold. *その他 [#ud02f636] **良い例 [#l8786e73] ***ほげほげパターン [#j08733ee] -Semantic Versioning versus Breaking Changes: A Study of the Maven Repository --To the best of our knowledge, our work is the first systematic study of versioning principles in a large collection of Java libraries. However, several case studies on backward compatible and incompatible changes in public interfaces as appearing in these libraries have been performed [2], [7], [9]. For instance, Cossette et al. [2] investigate binary incompatibilities introduced in five different libraries and aim to detect the correct adaptations to upgrade to the newer version of the library. Similarly, Dig et al. [9] investigate binary incompatibilities in five other libraries and conclude that most of the backward incompatible API changes are behavior-preserving refactorings. Dietrich et al. [7] have performed an empirical study into evolution problems caused by library upgrades. They manually detect different kinds of source and binary incompatibilities, and conclude that although incompatibility issues do occur in practice, the selected set of issues does not appear very often. --Another area of active research is to automatically detect refactorings based on changes in public interfaces [3], [4], [8]. The idea behind these approaches is that these refactorings can automatically be “replayed” to update to a newer version of a library. This way, an adaptation layer between the old and the new version of the library can automatically be created, thus shielding the system using that library from backward incompatible changes. --While our work investigates backward incompatibilities for given version string changes, Bauml et al. [1] take the opposite approach, in the sense that they propose a method to generate version number changes based on changes in OSGi bundles. A comparable approach in the Maven repository would be to create a plugin that automatically determines the correct subsequent version number based on backward incompatibilities and the amount of new functionality present in the new release as compared to the previous one. ***網羅して違いを言うパターン [#le1fcd41] --Evanco [7] also argued that introducing class size as an additional independent variable in a fault prediction model can result in misspecified models that lack internal consistency. --Zhou et al. [25] also addressed the effect of class size on the association between OO metrics and fault-proneness. The authors employed the Sobel test, which examines the significance of the mediation effect by calculating the standard error of X Y , evaluating the null hypothesis that the true effect of a mediating variable is zero. Their results showed that size has a mediation effect on the relationship between OO metrics and fault-proneness. Positively, the Sobel test is known to perform better (in terms of power) than the casual-steps approach [8]; however, there are a number of issues with this test, particularly its weakness due to the assumption of normality of sampling. In most cases the sampling distribution of X ! Y tends to be asymmetric, with high skewness and kurtosis [4, 18] (In our analysis of one of the two datasets used in [25] (originally generated in [5]), we found that the data were not normally distributed (see Section 3). In addition, the Sobel test has also been shown to score low in power and Type I error control [18]. -Revisiting the size effect in software fault prediction models ---Evanco [7] also argued that introducing class size as an additional independent variable in a fault prediction model can result in misspecified models that lack internal consistency. ---Zhou et al. [25] also addressed the effect of class size on the association between OO metrics and fault-proneness. The authors employed the Sobel test, which examines the significance of the mediation effect by calculating the standard error of X Y , evaluating the null hypothesis that the true effect of a mediating variable is zero. Their results showed that size has a mediation effect on the relationship between OO metrics and fault-proneness. Positively, the Sobel test is known to perform better (in terms of power) than the casual-steps approach [8]; however, there are a number of issues with this test, particularly its weakness due to the assumption of normality of sampling. In most cases the sampling distribution of X ! Y tends to be asymmetric, with high skewness and kurtosis [4, 18] (In our analysis of one of the two datasets used in [25] (originally generated in [5]), we found that the data were not normally distributed (see Section 3). In addition, the Sobel test has also been shown to score low in power and Type I error control [18]. -Note that all previous studies have investigated the size effect --Note that all previous studies have investigated the size effect from only one perspective, the indirect e!ect of size; that is, where size is a third variable and that the relationship between the metrics and fault-proneness is transmitted through this variable. As such, size is in this case a mediator variable, and such a variable is analysed using a mediation analysis techniques (i.e., studying the indirect e!ect).
