5 ms·
I know this post is about the "best" comments, but that doesn't necessarily indicate quality. I'm not sure what might be a good algorithm to judge quality, but
by physcab 17y ago
I know this post is about the "best" comments, but that doesn't necessarily indicate quality. I'm not sure what might be a good algorithm to judge quality, but popularity is most certainly not it. To all you natural language processing nerds out there...maybe there is a way of looking at word choice, with weights also given to amount of karma but relative to the other karma distributed to other comments.
- mkyc 17y agoThis is insane, not because you're wrong, but because popularity and quality should be equivalent for us. We have some of the very best natural language processing algorithms going over most of our posts - and several comments in this post are suggesting that they're failing miserably (in many important cases, I agree). This is a problem with the community, not with the software. One thing I dislike about the simple arrows is that they give no indication of on what basis posts should be upvoted. If, when clicking, we were forced to select from a menu vote types [+1 insightful, +1 clever, +1 informative, +1 antigroupthink, +0 I agree, +0 funny, -1 rude or troll], even if we did not track these, this would be giving people feedback on voting expectations and force them to pause for one second to put a bit more consideration into their vote.
- physcab 17y agoMaybe we should move to a 5-point rating system. A whole lot more data to work with!
- jongraehl 17y ago"About time... I can't be the only one who thinks Experts-Exchange is lame..." - I don't mean to impugn the author of that comment, but I think the amount of approval for it indicates that some people are voting on agreement more than quality.