Strengthen the server covenant #263
Conversation
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎
|
|
@Gargron @ClearlyClaire Any thoughts on this? It seems to me like a relatively straightforward requirement. For example, I don't think that a server could be considered moderated if it doesn't block many of the worst offenders and malware sources in lists like https://joinfediverse.wiki/FediBlock#Malware_&_co_Blocklist -- users who join such a server may well face serious harassment, especially if they are BIPOC, LGBTQ+, or otherwise likely to be targeted. |
|
I can imagine that Mastodon gGmbH might be reticent about specifically directing how a service provider perform the required "active moderation", how about: " Moderation must extend to abuse and inauthentic activity originating from third-party federated servers." Separately, and I'll work on adding this, it would be nice if the Covenant included links to community-accepted guides to administering and moderating content and conduct. Lastly, we are working on a recommendation to either
|
I'd like to hear from the maintainers before tweaking the language much further. IMO it is useful to spell out explicitly that it's a reasonable expectation for admins to limit or suspend servers which are routine abuse origins. We know that this is a very real risk for users joining the network -- a Black or LGBTQ+ user joining a server that lacks a basic blocklist may be immediately targeted by known abuse servers, in ways which are invisible to other users on the network. So I'm in favor of being as explicit as possible in the covenant that it's expected that responsible server operators will suspend such servers. This, combined with taking action to remove servers that aren't really trying to do this, is a very simple thing that Mastodon can do to strengthen the core network of servers people are pointed to. Fewer severs with actual moderation > more servers with shitty moderation. |
Explicitly reference abuse of external origin. A server that has a code of conduct and a small number of users may never feel the need to moderate abusive behavior. But if such a server fails to suspend the worst routine origins of abuse in the fediverse, it ultimately fails its users. In practice, new users signing up, especially vulnerable folks, may be immediately targeted by abusive actors from other servers. Several instances of exactly this happening have been well-documented by now. Since this extent of moderation responsibility is non-intuitive and non-obvious, it is important to be explicit.
eloquence
force-pushed
the
strengthen-the-covenant
branch
from
08d9db8 to
3a32754
Compare
|
Rebased, switched to up-to-date "limit" language instead of "silence". |
|
@Gargron @ClearlyClaire Just a friendly ping - any kind of maintainer response for this change would be appreciated. A similar change was recently accepted for the Firefish server guidelines (which serve a similar purpose as the covenant): https://git.joinfirefish.org/firefish/joinfirefish/-/blob/main/SERVER_GUIDELINES.md |
Explicitly reference abuse of external origin. A server that has a code of conduct and a small number of users may never feel the need to moderate abusive behavior. But if such a server fails to suspend the worst routine origins of abuse in the fediverse, it ultimately fails its users.
In practice, new users signing up, especially vulnerable folks, may be immediately targeted by abusive actors from other servers. Several instances of exactly this happening have been well-documented by now.
Since the extent of moderation responsibility is non-intuitive and non-obvious, it is important to be explicit.
This is not intended as a sudden or dramatic change, but as formal documentation of practices that most (but not all) servers in the covenant already adopt.
The language proposed here, or similar language, would enable us to gradually improve the moderation baseline among the servers featured on the welcome portal, to mitigate the likelihood that users signing up via Mastodon's official welcome portal experience harassment and abuse of the worst kind.
Resolves #261