• subignition@fedia.io
    link
    fedilink
    arrow-up
    6
    ·
    3 months ago

    You could do something like send a list of user IDs who voted, and have the client do the filtering locally with its blocklist. It would consume more bandwidth instead of computational power but probably wouldn’t scale very well

    • SorteKaninA
      link
      fedilink
      English
      arrow-up
      4
      ·
      3 months ago

      Well you could do something like that but it wouldn’t affect sorting of the posts then. If you have blocked an instance and that instance has upvoted a post to 1000 votes so that it appears on the top of the All-feed, you’d see it at the top with 0 votes (or only the votes from the instances you haven’t blocked), which would be very strange.

      • subignition@fedia.io
        link
        fedilink
        arrow-up
        2
        ·
        3 months ago

        Would the client not be able to sort the received posts after filtering? Granted things could be fairly inconsistent from one page to the next. Or maybe something like “Score: 0 (blocked: 1k)” to indicate the proportion of activity being taken out of consideration?

        Definitely not confident that this is a good idea or anything btw, just spitballing

        • SorteKaninA
          link
          fedilink
          English
          arrow-up
          3
          ·
          3 months ago

          Would the client not be able to sort the received posts after filtering?

          Your client only gets a page of the sorting at a time. You would only be able to sort that single page, but to do an actual sorting, you’ll need more posts than just what appears on the front page.

          • subignition@fedia.io
            link
            fedilink
            arrow-up
            2
            ·
            3 months ago

            Right, that is what I meant by “inconsistent from one page to another”. The received posts would be in an unfiltered state on each page and so the order could be very different depending on the local blocklist, probably resulting in stuff like the top of page 2 having a higher score than the bottom of page 1 pretty frequently.