I'm working on a new filtering engine that I think will be like nothing you've seen yet.
Basic overview:
- There will be a list of filters you can browse
- You select the filter list(s) you are interested in
- The filter list will receive updates.
- Say a new political event happens and a new name floods the news. The list can be updated dynamically to include the new name without you having to do anything.
For those of you that are tech savvy:
- Filter lists will live on GitHub
- You can fork an existing filter list.
- You can create/host your own filter list for other Blorp users.
- You can open pull requests for existing filter lists.
- There will be some tools within Blorp to debug your filter list
Other notes:
- I don't plan on letting you build these filter list from within the app. They are far too complex to edit easily within the app.
- But that complexity means you can do far more than other apps
- Block a post
IF
it mentions a US state (e.g. "Texas") AND
it mentions "shooting"
- Block a post
IF
it mentions "ICE" AND
it mentions ("detention" OR
"deport")
- Block a post
IF
it's by a specific list of users AND
it mentions a specific word
- The goal is to create more collaboration between users and make it easier to cut through all the noise on Lemmy and PieFed
How are you currently using filtering in Blorp, Voyager, etc?
- What type of content are you currently trying to block?
- Are your filters very unique to you, or something that could be shared as a list?
I just block the lemmit.online and zerobytes.monster instances and the lululemmy.com/c/redditmirror and 50501.chat/c/Mirror communities. There are probably more but those appeared in my feed so I blocked them. It isn't enough to block the specific bot accounts as the communities will still appear in the community list. Since the instances are as far as I can see just reddit archives I just blocked the whole instance.
The lemmit bot includes
in every post and the zerobytes one
in every post. There seems to be no consistent message with these bots, so I'm not surre text filtering would be more effective than instance/community/user based blocking