NSF supports inferring software specifications from open-source repositories

July 01, 2015

The National Science Foundation is supporting a large collaborative project on inferring software specifications from open-source repositories by leveraging data and the collective expertise of the developer community. Specifications describe how code is meant to behave, but most code ships without them, which makes reliable analysis and reuse harder.

The project mines the patterns latent across millions of projects, combined with community knowledge, to recover specifications automatically and at scale.

This work is part of Analyzing Software at Scale, with Boa; see our overview of reading the world’s code at once. Details are on the NSF award page.