Rank removal candidates

[smoelius1]

Recording an idea from @stansw...

Is there some way to rank removal candidates?

If there were, it could provide a way to limit false positives, e.g., by only running Necissist on the top n% of the the removal candidates.

[smoelius1]

Follow-up ideas:

• How do existing mutation testing tools rank mutants?
• Could ML be applied here? There is an issue to output a checklist. If a user checks an item, it signals that the item was worth checking. Could this information serve as training data?

[smoelius]

Some relevant papers:

• MuRanker: a mutant ranking tool: "ranks mutants according to their predicted difficulty and complexity in being detected"
• MuRS: Mutant Ranking and Suppression using Identifier Templates: "[ranks] mutants based on historical developer feedback"
• Prioritizing Mutants to Guide Mutation Testing: "[selects] mutants based on their predicted TCAP [test completeness advancement probability]"