dikaios1517 on Nostr: There are a lot of different spam mitigation techniques employed by both relays and ...
There are a lot of different spam mitigation techniques employed by both relays and clients. Some draw more ire from the community than others as "anti-free-speech censorship."
Here are some examples:
- Relays may require a payment in order to post to them. Though this may be criticized by those who want everything for free, it is an effective way to reduce spam, and running a relay that is available to the public has real costs involved that the relay operator ought to be able to charge for their services. Additionally, most paid relays offer other benefits, too.
- Relays may require a specific amount of proof-of-work (PoW) be completed before accepting the note. I don't know of any relays currently employing this policy, though. This is likely due to the lack of clients that allow for adding PoW to your notes. A large amount of users interact with Nostr via their mobile devices and adding PoW to everything they do would quickly become taxing on their phone's battery.
- Relays may only accept notes from users on a white-list. This list might be a fixed set of users, such as a list of community members, or it might be a dynamic list that can adjust itself over time, such as a web-of-trust (WoT).
- Relays may implement IP blocking or rate-limiting to discourage spam. IP blocking and IP-based rate-limiting has the downside of potentially blocking people who are trying to maintain privacy through the use of a VPN, though. This is because they will be using an IP address assigned by their VPN that several other users might also be using, therefore appearing like spam due to the volume of notes coming from the same IP address.
- Content-aware spam blocking might be employed by either relay owners or clients. This examines the contents of notes that an npub is publishing, and if they are repeatedly publishing the same or a similar note in a given period of time, the relay or client will flag their npub as a spammer and no longer accept notes from that npub, or display their notes to users who have spam-blocking options turned on in the client.
- Follower reports may also be used by clients to filter out spammers. For instance, if three or more of the people you follow have reported an npub as a spammer, the client may filter their posts from your feed. This feature can be too easily gamed, however. All it takes is a few people you follow to decide they don't want any of their followers to see a user's posts any longer and they could maliciously report them out of spite, rather than legitimate complaints with the content of their notes. Of course, not all clients implement this method of filtering out spam, so it would be pretty obvious that this has happened if you use more than one client, which a lot of folks do.
As it happens, most of the above mentioned methods of filtering out spam are also effective at filtering out pornographic content, too. In addition, though, a lot of clients have implemented NIP-36. This allows users to flag content as "sensitive." Either the original poster or other users may add this flag. Additionally, many clients have added certain hashtags as being considered sensitive content by default. It is then up to each user to determine in their client settings if they want sensitive content to be hidden or shown.
Here are some examples:
- Relays may require a payment in order to post to them. Though this may be criticized by those who want everything for free, it is an effective way to reduce spam, and running a relay that is available to the public has real costs involved that the relay operator ought to be able to charge for their services. Additionally, most paid relays offer other benefits, too.
- Relays may require a specific amount of proof-of-work (PoW) be completed before accepting the note. I don't know of any relays currently employing this policy, though. This is likely due to the lack of clients that allow for adding PoW to your notes. A large amount of users interact with Nostr via their mobile devices and adding PoW to everything they do would quickly become taxing on their phone's battery.
- Relays may only accept notes from users on a white-list. This list might be a fixed set of users, such as a list of community members, or it might be a dynamic list that can adjust itself over time, such as a web-of-trust (WoT).
- Relays may implement IP blocking or rate-limiting to discourage spam. IP blocking and IP-based rate-limiting has the downside of potentially blocking people who are trying to maintain privacy through the use of a VPN, though. This is because they will be using an IP address assigned by their VPN that several other users might also be using, therefore appearing like spam due to the volume of notes coming from the same IP address.
- Content-aware spam blocking might be employed by either relay owners or clients. This examines the contents of notes that an npub is publishing, and if they are repeatedly publishing the same or a similar note in a given period of time, the relay or client will flag their npub as a spammer and no longer accept notes from that npub, or display their notes to users who have spam-blocking options turned on in the client.
- Follower reports may also be used by clients to filter out spammers. For instance, if three or more of the people you follow have reported an npub as a spammer, the client may filter their posts from your feed. This feature can be too easily gamed, however. All it takes is a few people you follow to decide they don't want any of their followers to see a user's posts any longer and they could maliciously report them out of spite, rather than legitimate complaints with the content of their notes. Of course, not all clients implement this method of filtering out spam, so it would be pretty obvious that this has happened if you use more than one client, which a lot of folks do.
As it happens, most of the above mentioned methods of filtering out spam are also effective at filtering out pornographic content, too. In addition, though, a lot of clients have implemented NIP-36. This allows users to flag content as "sensitive." Either the original poster or other users may add this flag. Additionally, many clients have added certain hashtags as being considered sensitive content by default. It is then up to each user to determine in their client settings if they want sensitive content to be hidden or shown.