Skip to main content
Software Engineering

You are not logged in. Your edit will be placed in a queue until it is peer reviewed.

We welcome edits that make the post easier to understand and more valuable for readers. Because community members review edits, please try to make the post substantially better than how you found it, for example, by fixing grammar or adding additional resources and hyperlinks.

Required fields*

8
  • I guess the phrase "Not so important functionalities" is not so important (pun intended) for this question. The core part is whether an AI might be able to split down certain large commits into coherent smaller ones (ideally not just syntactically coherent, but functionality-wise). I am actually not sure whether this idea could work or whether there is a good reason why this will not work. Commented May 23 at 15:08
  • @DocBrown better now? Commented May 23 at 15:17
  • Test mapping was solved with mutation testing. The idea is to do coverage intersect with delta, then apply a compiling changes to the result, observing which tests fail. AI is probably unnecessary Commented May 23 at 15:27
  • @Basilevs an interesting point. I had been thinking of the commits as being strictly additive, meaning you'd be mutating to nothing rather than a different something. But thinking about it critically a commit doesn't have to be strictly additive. If we limit mutants to those in the commit history this might be plausible. Commented May 23 at 15:32
  • Fun fact: I asked Deep Seek which code is more important, the code for the breaks or the code for the rsdio, and got a long explanation why the brake system is definitely the more important system,. So I am not convinced an LLM cannot be used to classify certain features by importance. Commented Jul 21 at 20:00