Visit ShanghaiTech University | 中文
HOME > News and Events > Seminars
Detecting Architectural Debts
Date: 2015/7/21             Browse: 507

Speaker: Yuanfang Cai

Time: July 21st, 11:00-12:00 AM

Location: Room 220, Building 8, Yueyang Road Campus

Abstract:

Existing research has shown that files that were buggy in the past are also likely to be buggy in the future, implying that buggy files are seldom truly fixed. Our research has revealed that, in most cases, buggy files seldom exist alone; hundreds of buggy files can be connected by just a few architectural “roots”, created by implementation flaws that introduce unhealthy relations between files. These relations in fact cause groups of files to be buggy. Hence, it is impossible to reduce error or change rates without first identifying and then fixing the flaws that cause errors to propagate. Our objective is to locate these problematic files, and to identify the design flaws in them that cause bugs to arise, propagate, and persist.

In this talk, we introduce the key architecture model and a root detection algorithm for this purpose. We have validated our approach on dozens of open-source and commercial projects. Despite their widely varying nature, the detected roots have been shown to have significant and persistent impact on software maintenance costs. If not fixed, these roots accumulate maintenance costs just as debts accumulate interest. Our approach has already been adopted in 4 major international software organizations. Instead of examining hundreds of defective files in isolation, our collaborators now only need to examine a few architecture roots, fixing numerous defects simultaneously by removing their structural flaws, thus providing substantial long-term savings in maintenance costs. 

Bio:

Dr. Yuanfang Cai is currently an Associate Professor at Drexel University, USA. In 2006, Dr. Cai received her Ph.D degree in Computer Science from the University of Virginia, and started her career at Drexel upon graduation. She was granted tenure and promoted to associate professor in 2013.

Dr. Cai’s research focuses on software design, software architecture, software evolution, and software economics. Her recent work investigates architecture issues that are the root cause of software defects. In the past 5 years, her research received 6 awards from the National Science Foundation (NSF) of the US, with a total about 3 million dollars. Dr. Cai is currently serving on program committees and organizing committees for multiple top conferences in the area of software engineering, including ICSE, FSE, and ASE, and has been a reviewer for top journals in SE, including TOSEM and TSE.

                                                                                                                       

SIST-Seminar 15031