Identifying architectural hot spots by mining communication archives
|has title::Look who is talking: Identifying architectural hot spots by mining communication archives|
|Master:||project within::Software Engineering|
|Student name:||student name::Hanna Kollo|
|Second reader:||has second reader::Patricia Lago|
Melvin Conway postulated that the organisational strucuture of a softare development company is reflected in their products' software architecture (Conway's law). But how can one verify if this law holds for open-source software projects where developers are geographically distributed.
One possibility is to in investigate communication patterns between developers (by mining discussion lists, email archives, chat forums, etc.) and compare them to the architecture of the system. The investigation will reveal if communication takes place between developers to form tightly knit 'communication clusters', or communication is lose no such clusters exists. More importantly, it is important to verify if we can draw any parallers with the architecture of the product and identify any 'hot spots' that needs to be reviewed.
Requirements: Knowledge of any programming language (Java, Python, Perl) and some basic understanding of concepts in statistics, data mining, or artificial intelligence.