Robert Dyer, Computer Science PhD student advised by Dr. Hridesh Rajan, will be presenting a paper in the Technical Research Track at the International Conference on Software Engineering (ICSE) 2013, this May in San Francisco. His work with the Boa project attracted the attention of the CS Awards Committee, who has awarded Dyer the 2013 Tom Miller Graduate Fellowship. Students are nominated for the award by their faculty advisor. "I think that this research on the Boa project has the potential to change the way that software engineering researchers write code for mining software repositories with much ease, shorter learning curve, and higher reusability in term of libraries and existing code for common tasks. So Robert's presentation at ICSE will definitely help raise awareness in this simpler and more scalable research methodology. Eventually we are hoping that Boa will help propel a larger and more ambitious line of data-intensive scientific discovery in this area fueling reproducible experiments and enabling better collaboration in scientific discovery in this area." says Rajan.
The Boa project is a virtual laboratory for data-intensive research on open-source software development. Boa provides better abstractions for easily and efficiently mining ultra-large-scale software repositories. Data mining of large data sets - big data - requires significant knowledge on the part of the researcher in how to access and mine data, how to implement the infrastructrure required for data mining, not to mention the time required to perform the search and conduct the analysis needed to make decisions. Researchers in academic settings may not have the infrastructure. Researchers outside of computing specialties may not have the expertise. Boa lowers the barriers to entry for researchers seeking to make use of the data available in ultra-large software repositories, such as SourceForge, GitHub, and Google Code.
"Right now the Boa infrastructure includes one of the largest open source software repositories in the world," says Rajan, Associate Professor and Director of the Laboratory for Software Design. "Boa aims to be for open-source related research as Mathematica is to numerical computing, R is for statistical computing, and Verilog and VHDL is for hardware description. The main goal of Boa is to empower scientists and engineers interested in mining the wealth of software-related information stored in ultra-large software repositories by providing them with intuitve methods to efficiently answer research questions about trillions of lines of source code that belong to millions of projects, crafted and maintained over decades."
Dyer is an experienced student whose previous work includes Ptolemy and Nu, both projects directed by Rajan in the Laboratory for Software Design. Ptolemy provides the benefits of separation of cross-cutting concerns while maintaining the separate type-checking, compilation and reasoning properties of standard object-oriented languages. Dyer helped develop the research compiler (based on the JastAdd compiler framework) as well as performed large-scale empirical studies on the language. The Nu project is an aspect-oriented intermediate language. Nu adds two simple constructs to object-oriented programs and is capable of emulating most high-level constructs of languages like AspectJ, Eos, etc."
The Tom Miller Fellowship was established in 2012 by alumni and External Advisory Board member Tom Miller, to support and encourage CS graduate students to demonstrate excellence in research by pursuing opportunities to present their work at top tier international conferences and to publish in top research journals.