search

SOBotics / Nursery

A place to suggest and collaborate on new bot ideas.
0 stars 0 forks source link

Bot to suggest duplicates to common beginner questions #12

Open ztane opened 5 years ago

ztane commented 5 years ago

General topic: questions Requires feedback: yes Problem, the bot is supposed to solve:
In each tag there are common duplicate questions that share common code. Oftentimes it is even the same puzzling program that keeps asked over and over again, or just with minor variations. This is especially evident in the C tag. See for example this one that got 2 answers before it was hammered. A simple Google search quickly found 5+ duplicates with the exact same program structure.

Ideas for implementation: This is a bit tricky, it would require a language-specific removal of insignificant whitespace and possibly renaming variables and normalizing strings, then matching the code blocks against a database of taught duplicates, or list of common dupes from SEDE, for example by using simple Levenshtein distance.

ztane commented 5 years ago

And the more duplicates get closed solely with the canonical as a target, the more teaching material there is.