Our vision is a collection of 10,000 Java projects, each of which builds and runs, and for which popular tools succeed and have cached outputs. This National Java Resource will lower the barrier to implementation of new tools, speed up research, and ultimately help advance research frontiers. In particular, the National Java Resource will enable tools that take advantage of Big Code in such areas as code synthesis, error repair, and program understanding. Ideas for the National Java Resource include easy search for projects with desired characteristics, examples of how to run a new tool on the entire collection, and a web interface that enables anyone to run experiments.
Workshop goal: What do researchers need from a National Java Resource to make progress on their tools? A common road block is that existing collections of Java code are either small, without ability to build and run, or both. The main goals of the workshops are to:
- discuss the list of tools that researchers commonly use as building blocks for their own tools,
- debate what features of the National Java Resource that researchers would like to see, and
- see how an early prototype of the National Java Resource works.
We hope that several of the workshop participants will sign up to be early users and evaluators of our infrastructure.
A glimpse of the capabilities of an early prototype of a National Java Resource.
Mon 23 OctDisplayed time zone: Tijuana, Baja California change
08:30 - 10:00
|Welcome and Overview of NJR
|Benchmarking for Program Analysis
Cristina Cifuentes Oracle Labs, Karim Ali University of Alberta, Yannis Smaragdakis University of AthensFile Attached
|Hundreds of Projects and Tens of Thousands of Builds: Evaluating Dynamic Testing and Analysis Tools in the Cloud
Jonathan Bell George Mason UniversityFile Attached
|XCorpus, An Executable Corpus of Java Programs
Jens Dietrich Massey UniversityFile Attached
10:30 - 12:00
|Graph Systems for Statically Analyzing Hundreds of Projects at the Same Time
Harry Xu University of California, IrvineFile Attached
|Automated Benchmark Management
Eric Bodden Heinz Nixdorf Institut, Paderborn University and Fraunhofer IEM, Ben Hermann University of PaderbornFile Attached
|Tool Support for Energy-Aware Programming
Yu David Liu State University of New York (SUNY) BinghamtonFile Attached
|Discovering Empirical and Theoretical Analysis Properties from Lots of Real Executions
Michael D. Bond Ohio State University