Implement "harmful" and "harmless" super-categories - Githubissues

bettersg / checkMate

GNU General Public License v3.0

5 stars 1 forks source link

Implement "harmful" and "harmless" super-categories #317

Closed sarge1989 closed 3 months ago

sarge1989 commented 4 months ago

Problem

Sometimes, final outcome is unsure, due to voting category overlaps. For instance, legitimate overlaps with info(5). In such cases, users have to wait very long for a reply, and reply will say unsure, which is a bad experience.

Broad Concept

To solve this, we will map categories to either harmful or harmless
- scam, illicit, untrue map to harmful
- legitimate, accurate, spam to harmless
- trivial remains
- satire remains
If half of voters have replied but final reply is unsure, we additionally check if harmful or harmless, and a high consensus (>maybe 80%?). If harmless, reply anyway.

sarge1989 commented 3 months ago

One interim solution to this problem is to just immediately show the voting results message when it's unsure.