Links: Global optimization page | Origins of Lithuanian city names

Me

Audris Mockus

is the Harlan Mills Chair Professor of Digital Archeology in the Department of Electrical Engineering and Computer Science of the University of Tennessee, Knoxville. Min H. Kao Building, Rm 613, 1520 Middle Drive, Knoxville, TN 37996-2250, ph: +1 865 974 2265, fax: +1 865 974 5483, audris at utk.edu.
He also continues to work part-time as a consulting research scientist at Avaya Labs Research, ph: +1 908 696 5608, fax: +1 908 696 5402, audris at research.avayalabs.com.
Bibliography, curriculum vitae, research summary, and teaching summary.
Audris Mockus studies software developers' culture and behavior through the recovery, documentation, and analysis of digital remains, in other words, Digital Archaeology. These digital traces reflect projections of collective and individual activity. He reconstructs the reality from these projections by designing data mining methods to summarize and augment these digital traces, interactive visualization techniques to inspect, present, and control the behavior of teams and individuals, and statistical models and optimization techniques to understand the nature of individual and collective behavior.


Results from Digital Archeology include the ability to determine why the software is changed, how difficult each change is, to evaluate the impact of a software tool or process, and to predict risk that a change will break existing features. Digital Archeology also allows quantification of key features of a development process including  Open Source development process. Details are in the Apache server case study and the scripts used in the commonly cited ( times) Apache and Mozilla study.

Software development practice is experiencing a radical change driven by the open source movement, the business needs to move development to low-cost locations, the aging and renewal of core developers in legacy products, recruiting in fast growing Internet companies, and the turmoil of the economy causing unprecedented turnover in software projects. The investigations on transfer of work and the associated phenomena related to organizational change and the growth of software project competencies clarify the nature of these challenges. In particular, the relationship between the social and technical competencies is associated with the fraction of new participants who become long term contributors. Also, the initial environment a person encounters when joining a team or an organization affects motivation and long-term behavior.

Given the diversity of software projects and the creative nature of software development, the context may strongly affect project's sucess. Multi-company studies [ASM08,MNT09,CMRH09] took into account not only technical, but also social [HM03b,CMRH09] aspects of software development. Surprisingly, a software project with identical requirements may cost an order of magnitude more and require correspondingly more effort simply because of the differences in the nature of company's customer base [ASM08]. At the same time, many phenomena related to how developers make mistakes leading to software defects are similar in diverse projects and companies [M10,CMRH09,MNT09].

Applications of Digital Archeology in the context of globally distributed software development lead to tools for optimal to distribution of work, estimation and visualization of developer expertise (a demo for Mozilla code), quantification of distributed projects' lead time drivers. Ways to measure and improve the quality of software [MW00,IQ08] are used to determine if projects meet their quality targets, to guide strategic company decisions HMPQ10, and to demonstrate that the reliability of products meets the requirements of existing and prospective customers.

Earlier work includes Live Document web-based visualization and presentation technology to present and explore complex data and global optimization that helps to fit and test models associated with complex systems.

In an earlier work I analyzed spatio-temporal data including estimation of covariance function from aggregates with Layout Analysis and interactive aggregation techniques to display and explore such data.

audris@mockus.org