Comment on alien.top is a new level of Reddit crossposting spam

• rglullis@communick.news ⁨11⁩ ⁨months⁩ ago

  Why?

  source

  • uis@lemmy.world ⁨11⁩ ⁨months⁩ ago

    Why hard? Client needs to fetch all metadata needed for sorting for every post created during entire lemmy’s existance on every discovered lemmy instances, which depending on algo you are using, might include comments metadata. To aid client-side sorting you would need server-side filtering, which will limit data avaliable to sorting algo. For example client-side trending algo would not show old trending post because it was filtered out.

    So client-side sorting is basically running stripped version of instance without file hosting.

    source

    • rglullis@communick.news ⁨11⁩ ⁨months⁩ ago

      Client needs to fetch all metadata needed for sorting for every post created during entire lemmy’s existance on every discovered lemmy instances.

      Why would you need all data to build the frontpage? Why not just make a sliding window with the content from the last 24/48h?

      So client-side sorting is basically running stripped version of instance without file hosting.

      Even if that were true, how is that different, e.g, from any modern desktop email client?

      source

      • uis@lemmy.world ⁨11⁩ ⁨months⁩ ago

        Why would you need all data to build the frontpage? Why not just make a sliding window with the content from the last 24/48h?

        Exactly what I’m saying. To not be super resource-intensive, client-side sorting needs to be incomplete. As I said, if there is hypothetical post from 49 hours ago with 10k upvotes, you will not see it, but you will see one from 48 hours ago with 1k upvotes.

        Even if that were true, how is that different, e.g, from any modern desktop email client?

        Not much I guess. But what about mobile?

        source

        • -> View More Comments