-

2025-03-29 06:23:14

plasticpuddle@sharkey.world

Is there a way to use a universal keyword blocking filter for all your #fediverse accounts?

Instead of having to create one from scratch every time. Like instead of one for Lemmy and one for Mastodon and one for Pixelfed, you just use one filter?

2025-03-29 09:19:20

Jupiter Rowland

jupiter_rowland@hub.netzgemeinde.eu

@plasticpuddle No, and there never will.

For one, there is no way to sync these filters not only between server instances (unless you're on Hubzilla, (streams) or Forte), but even between instances of completely different servers.

Besides, not all filters are fully compatible with one another.

On Mastodon, you can define multiple individual filters for your whole account, and they also cover hiding posts behind a generated CW. On Hubzilla, (streams) and Forte, you have one filter for your entire channel with a whitelist and a blacklist, hiding posts behind a generated CW is done by a separate filter system, and in addition, you can have one individual filter with a whitelist and a blacklist for each of your contacts. It'd be quite an effort to translate back and forth between these.

On Mastodon, you can configure per filter whether the keywords in the list shall only be recognised as whole words or not. Hubzilla, (streams) and Forte use regular expressions. For one, you'd have to translate back and forth between one Hubzilla/(streams)/Forte filter with regex and multiple (and always the same) Mastodon filters. Besides, Hubzilla, (streams) and Forte can do a whole lot more with regex than Mastodon can do with its settings switches (match the beginning of words, match the end of words, make characters optional etc.).

Hubzilla, (streams) and Forte also have something called "filter syntax" (which is mutually exclusive with regex, though). For example, ?verb == Announce in a blacklist filters out boosts. Hubzilla needs this in a filter line, but it can be used both for the whole channel and for individual contacts. (streams) and Forte have a dedicated switch for this for individual contacts, but the line doesn't have to be translated to that switch because the filter syntax line still works. But many other Fediverse server applications don't even have the capability to filter boosts out, much less filter everything out that isn't a boost which is what happens when you enter that line into a whitelist.

If, instead of syncing between instances, you could actually set up one list outside these instances, that list would only cover the common bare minimum of filter features available in all Fediverse server applications. Say buh-bye to CWs created reader-side by a keyword list. Say buh-bye to regex. Say buh-bye to filtering boosts away. Say buh-bye to letting comments and/or PMs pass through unfiltered (available on Hubzilla, (streams) and Forte by using filter syntax).

Besides, as long as there is no decentralised Fediverse server application where you can host your filter list, you'd have to host it yourself. As in, either run a small Linux computer with no screen at your home 24/7 or pay for a Web server and, again, learn Linux and the command line. But any of this is moot because no Fediverse server application can pull in personal filter lists from external sources.

#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Filter #Filters #Mastodon #Friendica #Hubzilla #Streams #(streams) #Forte