Code Delay in Legacy Software Systems: Measurement, Models, and Statistical Strategies

这个研究计划针对的是一个巨大但基本上没有得到重视的问题的关键问题--计算机软件的腐朽。需要研究的三个关键问题包括：描述代码衰减，识别导致代码衰减的因素，以及评估减少代码衰减的效果。该研究团队代表着AT&amp；T和美国国家统计科学研究所之间的合作关系，将开发反映软件开发过程中关键行为和管理特征的代码衰减指数(衡量标准)。可控(例如，组织)因素的变化将改变指数，进而影响关键反应：成本、时间和质量。通过结合适当的行为和组织因素，以及代码特征，我们将开发技术来识别相关因素并评估它们对代码衰减的影响。研究团队是跨学科的，包括软件工程、行为问题、组织理论、管理和统计方面的专家。统计学--尤其是测量、建模、预测和实验--是解决代码衰减问题的“粘合剂”和使能技术，也是将传统科学的范例应用到软件工程中的手段。将开发并使用统计策略来模拟和控制设计决策、软件架构和组织因素的影响。这些产品将成为管理工具的基础，以实现降低成本和开发间隔、提高软件质量和增强预测成本、间隔和质量的能力。

This research program is directed toward key problems regarding a massive, but largely unappreciated problem -- decay of computer software. Three key problems to be investigated include describing code decay, identifying factors that cause code decay, and evaluating effects of reductions in code decay. The research team, which represents a partnership between AT&T and the National Institute of Statistical Sciences, will develop indices (measures) of code decay that reflect key behavioral and managerial characteristics in the software development process. Changes in controllable (for example, organizational) factors will change the indices, which in turn impact key responses: cost, time and quality. By incorporating adequate behavioral and organizational factors, as well as code characteristics, we will develop techniques to identify relevant factors and evaluate their effects on code decay. The research team is cross-disciplinary, including experts in software engineering, behavioral issues, organizational theory, management and statistics. Statistics -- especially measurement, modeling, prediction, and experimentation -- serves as the `glue` and enabling technology for progress on problems of code decay, and as the means to apply paradigms from the traditional sciences to software engineering. Statistic strategies will be developed and used to model and control the effects of design decisions, software architecture and organizational factors. The products will be the basis for management tools for achieving reductions in cost and development interval, increases in software quality and enhanced capability to predict cost, interval and quality.