in Link Post

Automated Bug Triaging, or: How to Document Your Method


IBM Research, India:

In this research, we propose a novel bug report representation algorithm using an attention based deep bidirectional recurrent neural network (DBRNN-A) model that learns a syntactic and semantic feature from long word sequences in an unsupervised manner.

This is the support site for the research paper which is linked to at the top of the page. The actual research is somewhat interesting, but the big positive takeaway here is the content of this site. As noted in the abstract:

Another major contribution is to make this research reproducible by making the source code available and creating a public benchmark dataset of bug reports from three open source bug tracking system: Google Chromium, Mozilla Core, and Mozilla Firefox.

Sure enough, this site contains the complete and (as far as I can tell) very well documented source code for the project. Plus links to the entire data set. Computing power notwithstanding, it should be trivial for anyone to reproduce the results shown in the paper. That’s amazing. On the other hand: it shouldn’t be amazing, it should be normal. Small steps, though.