Debugging with intelligence via probabilistic inference

blog.acolyer.org

155 points·godelmachine··5 comments

I can't find a link to the tool to try the debugging approach on my own code. Have I missed it or there is none usable?

Since it's an academic project, I doubt that they published the code. You could try asking the first author [1], but don't expect too much. Academic code tends to focus on getting a paper published, and code quality and usability tends to suffer. I wouldn't be surprised if they didn't use source control and don't have it in a runnable state right now.

[1] https://sites.google.com/site/pyzhaogui/

This is changing though. At least in Europe there is a growing pressure for reproducibility and openness of data and implementations. I hope that in a recent future publicly funded content will have to be released, no matter how experimental or inflexible. Unless there are good reasons to keep it private (e.g. 0 days during embargo).

That depends on the field. Researchers are encouraged to publish artifacts with their paper in some fields to increase reproducibility. For example, most big programming language conferences (PLDI, POPL, OOPSLA etc.) let you submit your artifact, get it evaluated and award your paper with a badge noting you also have an available, working implementation.

Are there products that are already using probabilistic inference for debugging already? How mature is this approach?

Reply on news.ycombinator.com