Re: Old directions in social media.

Michael Thomas <mike@xxxxxxxx> · Wed, 6 Jan 2021 10:27:51 -0800



    On 1/6/21 10:18 AM, Rob Wilton
      (rwilton) wrote:

    
        I’m
            not really wanting to get dragged into this debate … but I
            think that it might be interesting to try and objectively
            compare the quality of the documents produced by WGs making
            use of github vs those just using email.
         
        One
            swallow does not a summer make, but I note that the 4 QUIC
            documents that are in IESG review all seem to be
            exceptionally polished and well written, and I understand
            that the QUIC WG has a strong Github based editing process. 
            Of course, there could be many factors as to why this is
            (e.g., maybe the editors are native English speakers, these
            documents probably have had more review and discussion) but
            it also seems plausible that easier collaborative
            editing/updates might be a factor in the high quality output
            from the WG.
         
        
    I'm fairly certain that it has far more to do with who is
      authoring/editing the documents than any tools. My favorite
      whipping boy rfc 8226 is terrible and woefully underspecified, and
      when that was pointed out in last call, it got aggressive push
      back. Tools are not going to change that sort of problem. What
      we'd need is AI's that can emulate the normative language and poke
      holes in it. Sometimes the I's are up to the task but too often
      they are not, or are ignored.

    
    Mike