Ultimately, in principal, what counts as 'spam' is a subjective judgement and should probably be treated as such. To me, that suggests an elegant reputation system.
I'm just thinking out loud here, so take the technical feasibility with a huge grain of salt: What if there was a mechanism for automatically muting anyone who 10% of my follows had already muted? - and that would cascade so that if I were automatically muting someone based on 10% of my follows (or explicitly muting), that would count towards the 10% threshold for any of my followers to also automatically mute, and so on. So, once a new spam bot popped up, it seem like it wouldn't be long at all before that spam bot was effectively eliminated for anyone with just a few follows via a cascading reputation system. BTW, the 10% threshold is just an arbitrary example value. Ultimately, It would probably be best to be adjustable on an individual user basis.
nostr:nevent1qqs0nwuu9fnex9r2jqu2aew8ar25js85msjrxckvk7e5v6tct68gzxgpz4mhxue69uhhyetvv9ujuerpd46hxtnfduhsyg89yuk7j99axqt4t3pehz8xjkdy8jwjveyrruync50fc7v6z6ss9upsgqqqqqqswrr3se
I like it. I'm not sure how well that much granular control would work at a user level, but it sounds like a reasonable approach for relays to take. It seems to me we need elegant solutions at both levels.
Yea I agree and would add that the granularity would be obscured to the user, who would be presented with some sane defaults and friendly interface options (which control the granularity under the hood).
Like when you enter a word into a "filter" setting field. You don't have to write the pattern-matching regular expression and detail which exact fields of a note you want to match on... You just write a normal word and hit save.
Or in many cases, you choose a client that has a lot of spam protection built in and you don't even know it.
I would expect to see dozens of clients appearing that all handle WoT in different ways without burdening the user with too many options.
Agreed. This is inevitable. I expect it will also become more difficult to discern which of those clients are 'protecting' you from more than just spam.
Agreed, but that's also why Nostr is a phase change and not an incremental improvement.
You can always exit a client with no loss of data or network.
This is what Amethyst was/is doing and people created lots of FUD around that approach as being evil cenzorship.
I have no experience with Amethyst (no android phone). But, any solution will have its critics. As for the haters, 'censorship' falls flat to me since all content would still be unmuted to anyone who chose a reputation threshold of 0%. Also, individual follow lists should probably override this mechanism on that individual's basis as well (i.e. someone you explicitly follow should not get automatically muted). So, IMHO, either of these available options would necessarily invalidate the 'censorship' tag.
Oddbean new post about login | logout