#[0] not that i want to take up the lead developer's time too much here, but i'm just wanting to know if maybe a naive Bayesian classifier could separate the cream from the crop? most of spam is eliminated with such a classifier. you're using degrees of separation in your client. but we've got decades of computationally inexpesive Bayesian classification to go on here. just a little curious if this was thought of?

Reply to this note

Please Login to reply.

Discussion

addition: what your client does is definitely working but might shut out newcomers. if other clients copied this approach, no new people would be seen, as i understand this algo.

in a moment of forgetfulness, i should have added that this is how most email spam filters work, and they're pretty good at it. i feel this technique may be underutilised.

Thanks. We can consider it in the future or even better if you do a PR and it could an alternative option :)

haha i know what PR means: a whole lot of dev time with no guarantee of a reward at the end of it. you can reject that. how do i know my time isn't wasted? 😆

i write good code. former employers can attest. but submitting PRs to various projects seems like a fool's game to me.

Iris is an open-source project. You're not paying me for replying either 😄

well, you know, even the GNU licence allows you to charge for things

with that said, there's some goal you have on the long term, i assume, that isn't altruistic.