1. 14
    Promise Pipelining in Cap'n Proto RPC ☶ distributed networking programming kentonv.github.io
    BruceM avatar via BruceM 4 years ago | cached | 5 comments

He also wrote about this on his blog in December: http://kentonv.github.io/capnproto/news/2013-12-13-promise-pipelining-capnproto-vs-ice.html

  1.  

  2. 6
    kragen avatar kragen edited 4 years ago | link

    This is one of the technologies planned to support server federation in Sandstorm.io, one of a group of emerging decentralized application platforms that are our only hope to escape the disastrously centralized web we seem to be blundering into. It’s based on the promise-pipelining technique from E and CapTP and described in more depth in MarkM’s dissertation, though it looks like Kenton’s only implemented the 2-vat variety of promise pipelining so far — which is fine, since as far as I can remember they never got the 3-vat variety running in E, either.

    Each of these individual decentralized application platforms by itself is not going to be enough; most of them will fail. But together they comprise our only hope for resurrecting some semblance of liberal democracy.

    1. 2
      JohnCarter avatar JohnCarter 4 years ago | link

      only hope to escape the disastrously centralized web we seem to be blundering into. our only hope for resurrecting some semblance of liberal democracy.

      I think you are deluded about cause and effect here.

      The long standing lack of liberal democracy, and the presence of economic, property, cultural and political systems driven more by “1 dollar == 1 vote” than democratic ideals, has firmly driven us towards the disastrously centralized web.

      No blundering, only clear intent, involved.

      From what I can read of history, I suspect your phrase

      hope for resurrecting some semblance

      would be more accurate if permuted.

      resurrecting hope for some semblance

      1. 2
        kragen avatar kragen 4 years ago | link

        The world is a great deal more complex and varied than your comment suggests.

      2. 2
        kragen avatar kragen 4 years ago | link

        I think Irmin and OpenMirage in general are another such crucial attempt which happens to be on lobste.rs today.

      3. 1
        JohnCarter avatar JohnCarter 4 years ago | link

        If I’m reading the docs / code correctly….

        The BIG gotcha with Cap'n Proto is it’s big selling feature is also it’s BIG downside….

        This benchmark is, of course, unfair. It is only measuring the time to encode and decode a message in memory. Cap’n Proto gets a perfect score because there is no encoding/decoding step. The Cap’n Proto encoding is appropriate both as a data interchange format and an in-memory

        ie. It provides accessors for each field… but doesn’t pack/unpack into/from vanilla C/C++ struct / POD.

        ie. If I have a vanilla C/C++ struct/POD (Plain Old Data) and or a vanilla function declaration, I cannot shift the definition to Capn'proto IDL, compile, get back exactly the same declaration.

        I cannot hand the POD to the framework to pack/unpack.

        I cannot just invoke / implement the function and expect the call to be serialized and routed and invoked on the far side with the data passed as POD parameters.