[ih] Today???s Internet Still Relies on an ARPANET-Era Protocol: The Request for Comments (Steve Crocker)

[Toerless Eckert]

On Sat, Aug 08, 2020 at 12:00:52AM -0700, Jack Haverty via Internet-history wrote:
> I just looked at a few recent RFCs, and they look amazingly like they
> did 50 years ago.  Like they came out of a typewriter.  I know there's
> been some evolution to move beyond pure ascii text, but still they look
> very much as they did in the 70s.  Very traditional and nostalgic, but
> fifty years has provided a lot of time for new approaches to mature.

Indeed. And the tooling RFC-editor now has for better graphics is IMHO
still challenging when you try to create better graphics for e.g.: network
diagrams or the like. And for some things such as showing protocol
headers, ASCII graphics is IMHO ideal for human consumption. Not so much
for formal verification though.

> There are many implementations now of collaboration, but one I find
> interesting is github.  It is used traditionally for coding projects,
> but it does a pretty nice job of handling documentation too, including
> discussions, revisions, etc.

Except that we now (IMHO) struggle with a rift in the community: Those
who breathe github and expect that all process of work has to be looked
up on github and those like me who think that rfcdiff is just lightyears
ahead of a diff i can get on github and would therefore rather prefer
to have all versions in datatracker. Then again issue tracking is nice
in github - as you also mention below.

> One example of this is in the "PDP-10/ITS" area on github --
> https://github.com/PDP-10/its/issues -- which contains "issues" (an
> "issue" is somewhat analogous to an RFC), coupled with an ongoing
> repository of the ensuing discussions about that issue.   People can
> engage in discussions directly on github, or interface through email.   
> Text, images, video, etc.  Oh, lots of code is there too.    I can see
> RFCs simply being "issues" in github, which would capture and make
> available an evolving archive of the ensuing comments and discussions.

Yes. And now we've got scatter/gather storage of discussions...

> This is only one example of what might be called modern collaboration
> practice on the Internet.  All the IETF would have to do is publish each
> new (and even old) RFC as an "issue" in a github area, and publicize how
> to submit your comments.
> 
> But I know, it's hard to change tradition....   Just an idea.

https://tools.ietf.org/html/draft-ietf-git-using-github-06

Cheers
    Toerless

> /Jack
> 
> On 8/7/20 11:10 PM, Toerless Eckert via Internet-history wrote:
> > I think that emails to IETF mailing lists are closest to the original spirit
> > of "requests for comments" these days.
> >
> > Maybe we could introduce a new email message header that IETF mailman would insert,
> > something like "RFC-ID: RFC.MSG.sender.date.serial.list" (or the like).
> >
> > This would give each mail to an IETF mailing list a unique "RFC" identification,
> > and thereby elevating it to an RFC. While primarily meant as a honorary identifier
> > to remind of the original spirit of the term, this could even be useful because
> > those IDs could be used as references if we also come up with a URI/URL scheme for
> > them (worst case an IETF mailman URL). [ Not sure if there are even URLs for
> > standard Message-ID's on mailman. ]
> >
> > Other than that, it would be lovely if we would have better recommendations
> > for changelogs in drafts to capture the discussion. Today this is left to
> > authors, and WG chairs typically don't dare to ask for more if authors don't
> > want to do it.
> >
> > Cheers
> >     Toerless
> >
> > On Fri, Aug 07, 2020 at 03:49:39PM -0700, Dave Crocker via Internet-history wrote:
> >> On 8/7/2020 3:16 PM, Vint Cerf via Internet-history wrote:
> >>> then came Internet Drafts which highlighted conversation again.
> >> Even I-Ds tend to be used in a more formal way, now, making them more
> >> representative of RFC development snapshots than of comment indicators,
> >> which are handled in email and meetings.
> >>
> >> There is, occasionally, an I-D done as a comment on other work -- with no
> >> intent for RFC publication -- but that's extremely rare.
> >>
> >> d/
> >>
> >> -- 
> >> Dave Crocker
> >> Brandenburg InternetWorking
> >> bbiw.net
> >> -- 
> >> Internet-history mailing list
> >> Internet-history at elists.isoc.org
> >> https://elists.isoc.org/mailman/listinfo/internet-history
> 
> -- 
> Internet-history mailing list
> Internet-history at elists.isoc.org
> https://elists.isoc.org/mailman/listinfo/internet-history

-- 
---
tte at cs.fau.de