I recently finished the episode of The Verge’s podcast #Decoder with the interview to Bluesky’s CEO and it seems a quite interesting project. At the beginning I wasn’t looking really into it because of their choice of using a new protocol instead of the existing ActivityPub, but after listening to her and the reasons behind this choice maybe I’ll give them a chance.

What do you think? Do you use it alongside with the fediverse?

  • @SorteKanin@feddit.dk
    link
    fedilink
    English
    88 months ago

    I don’t know much about their protocol

    As far as I understand, Bluesky is basically a central authority in their protocol. I wouldn’t really call that better than ActivityPub.

    • @Plopp@lemmy.world
      link
      fedilink
      English
      08 months ago

      Well, there are many different aspects to take into account. I was thinking more of how inefficient AP is when it comes to system resources and network usage, and some other things I can’t remember that made me go “yikes” when I read it. Also how it’s used for things the protocol doesn’t really have support for, so devs make their own solutions that are now part of the AP Fediverse even though the protocol itself, that is the backbone of the thing in question, doesn’t support the things that is a part of the thing. It seems a big mess in many ways, and I believe that Bluesky doesn’t have those issues.

      • @SorteKanin@feddit.dk
        link
        fedilink
        English
        108 months ago

        Can you be more specific? How is AP inefficient? What are the nonstandard extensions that devs have made?

        • @Plopp@lemmy.world
          link
          fedilink
          English
          08 months ago

          It’s been quite a while since I read about the inefficiency. I think it had something to do with CPU load, and that it’s unnecessarily “chatty” in some ways that causes servers to use unnecessarily large amounts of data. And the extensions had to do with different types of services, where the AP spec is best suited for one type of service (like maybe micro-blogs iirc), and others have to use the spec in weird ways or add things on top of it to implement other features that are important for those other types of services, like more forum-esque type things like Lemmy. Don’t remember exactly what they were, but one thing I read last week was that guy who had to shut his AP project down because he used a method of fetching data, that Mastodon (or whichever service it was) uses but isn’t part of the AP spec, and poorly documented, so he implemented it wrong which had horrendous consequences for him, but that’s a different story.