1. 9

Pump.io is a protocol and a server for a federated social web.


  2. 4

    It’s great to see the community stepping up and making things happen, now that Evan has handed over the project.

    2.0.0 is an upgrade of several of the dependencies, but not all. 3.0.0 will happen pretty soon with more changes that allow all dependencies to be up-to-date (IIUC and IIRC). After that it’s probably time to start migrating to W3C’s ActivityPub, which is inspired by pump.io and based on ActivityStreams 2.0.

    pump.io and also OStatus/GNU Social (a.k.a. StatusNet, laconi.ca, known from e.g. identi.ca which is nowadays running pump.io) are based on ActivityStreams 1.0.

    1. 2

      the community

      Well… “the community” is one guy, right?

      1. 1

        Strugee is doing the bulk of the actual commits, but there’s more than one person at the weekly IRC meetings, and others are e.g. going to adopt existing servers.

      2. 2

        Are these projects interoperable (pump.io and Gnu/Social)?

        1. 2

          Sadly, no.

          OStatus federates only public messages, pump.io has authorization, so you can restrict access to certain people. Pump.io has a more lightweight propagation mechanism.

          OStatus is (very roughly) AS1+Atom+PuSH+Salmon, pump.io is (very roughly) AS1+some envelope with e.g. authorization+auth pingback+REST.

          1. 1

            Too bad, it would be nice to have a way to communicate between the 2 (even if only partially, like only public messages). This way given mercer’s comment, I’m more inclined towards using Gnu/Social, given the set of projects that are compatible.

            1. 2

              Yeah, unfortunately pump.io had to be a breaking change, as the purpose was to untangle the protocol stack so that it could be a simpler and slimmer service. Evan burned the disk packs and even switched implementation language, because he felt he couldn’t do it quickly with the codebase he had accumulated.

              My hope is that ActivityPub, being a W3C standard, could get some people interested in a GS implementation, even if only for public messages. That would mend the gap.

          2. 2

            GS and Pump aren’t, but GS is compatible with several other projects that also implement OStatus: Friendica, Mastodon, Hubzilla, partly Diaspora (with a bounty going on for full OStatus support!).

            1. 2

              It’s cool to see improved OStatus getting started in Diaspora.

              Friendica (and Hubzilla?) also speaks pump.io.

        2. 3

          Congrats on the release! I like Evan’s work on the activity stream spec.

          1. 1

            I’m not directly involved, just a big fan, but thanks anyway! I’m happy it’s moving forward.

          2. 1

            I was debating wether to label this release or not, as it’s a beta. Now I’ve learned that the label is appropriate. Thanks, all!

            1. 1

              I wish pump had never happened and instead GnuSocial/StatusNet would have been extended. Nevertheless, I’m happy that development is going forward.

              1. 2

                That would have meant identi.ca being dead instead of the state it is in today. Even if the pumpocalypse drove many users away, the alternative was worse.

                1. 1

                  Why was identi.ca not viable with GS?

                  1. 1

                    “I think I can run identi.ca for somewhere between 10 and 25% of the current operating costs (which are considerable, for me).”


                    Later statements from Evan indicate that this estimation basically held true.

                    1. 1

                      (still before the transition)

                      “Identica right now costs around $6K/month. With pump.io, I think I can get it down to under $500/month.”