[ih] Host-Host vs NCP (Was Re: Confusion in the RFCs)

[Steve Crocker]

John,

Thanks!  I don't think I ever saw that survey.  October 1974 was an
interesting time for me.  I had finished three years at DARPA, where my
primary focus was on AI, not networking.  I had gotten married in September
and was on an extended vacation/honeymoon and cross country to move to
ISI.  We were on the road for about six weeks, from mid September to early
November.  And when I arrived at ISI, I plunged directly into formal
methods.

The division of Network Control Programs into big and little and their
correlation to the quality of existing interprocess communication
facilities seems exactly right to me.

Steve


On Fri, Sep 5, 2025 at 10:59 AM John Day via Internet-history <
internet-history at elists.isoc.org> wrote:

> Well, a little earlier but indicates it was in wide use prior to Oct 74.
>
> Postel, J. B., Survey of Network Control Programs in the ARPA Computer
> Network, MITRE Technical Report #6722, October 1974, Revision 1, 89 pp.
>
> BTW, this is a very important document. There is an important result in it
> even though Jon didn’t note it .
>
> Reading it is clear that NCPs fall into two categories: big and little.
> The difference was those OSs with good IPC had small NCP implementations.
>
>
>
> > On Sep 5, 2025, at 10:45, John Day via Internet-history <
> internet-history at elists.isoc.org> wrote:
> >
> > So, you are saying that NCP didn’t come into use until 74-76? First TCP
> spec I saw was in mid-75.
> >
> > I can’t cite anything off the top of my head but I think it was in use
> well before that.
> > Someone else will remember.
> > John
> >
> >> On Sep 5, 2025, at 10:40, Noel Chiappa via Internet-history <
> internet-history at elists.isoc.org> wrote:
> >>
> >>> From: Steve Crocker
> >>
> >>> Somewhere along the way, people repurposed "NCP" to mean Network
> >>> Control *Protocol*
> >>
> >> When there was only one protocol stack running on the ARPANET (HHP et
> al), it
> >> didn't need a name. When the second one (TCP/IP) appeared, we needed
> names
> >> for each, and we started using 'NCP' as the name for the first one
> ('NCP' was
> >> sort of in use already to refer to that collection of things, as the
> name for
> >> the software that implemented it), and backronymed it to the more
> >> sensible/appropriate 'Network Control Protocol'.
> >>
> >> Fun fact: the CHAOS people at MIT retained, for some years, 'NCP', and
> the
> >> old expansion of it, as the name for the CHAOS protocol implementation;
> for
> >> example, see:
> >>
> >> https://github.com/bill790/chaos/blob/master/ncp.c
> >>
> >>      Noel
> >> --
> >> Internet-history mailing list
> >> Internet-history at elists.isoc.org
> >> https://elists.isoc.org/mailman/listinfo/internet-history
> >> -
> >> Unsubscribe:
> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history
> >
> > --
> > Internet-history mailing list
> > Internet-history at elists.isoc.org
> > https://elists.isoc.org/mailman/listinfo/internet-history
> > -
> > Unsubscribe:
> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history
>
> --
> Internet-history mailing list
> Internet-history at elists.isoc.org
> https://elists.isoc.org/mailman/listinfo/internet-history
> -
> Unsubscribe:
> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history
>


-- 
Sent by a Verified

sender