From dhc2 at dcrocker.net  Tue Jan  3 07:57:36 2017
From: dhc2 at dcrocker.net (Dave Crocker)
Date: Tue, 3 Jan 2017 07:57:36 -0800
Subject: [ih] Fwd: Reporter query on the history of greater-than in quoted
	replies
In-Reply-To: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
Message-ID: <cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>

Hi folks.

Just got this query.  I don't have any clear memory of how the usage got 
started.  Do any of you?

d/

-------- Forwarded Message --------
Subject: Reporter query on the history of greater-than in quoted replies
Date: Mon,  2 Jan 2017 12:43:04 -0800
From: Glenn Fleishman <glenn at glennf.com>

...
I'm a reporter with a deep background in typography, and recently wrote 
this piece about quotation marks (the curly kind) for The Atlantic: 
https://www.theatlantic.com/technology/archive/2016/12/quotation-mark-wars/511766/

In researching that story, I found that the > mark was the original 
quotation mark ? it dates back about two thousand years in that usage. 
This fascinated me, because of how it was revived in email and forums.

I read early RFCs and other documents, and was unable to find the origin 
of > for quoting a reply.
...


-- 

   Dave Crocker
   Brandenburg InternetWorking
   bbiw.net


From scott.brim at gmail.com  Tue Jan  3 08:38:48 2017
From: scott.brim at gmail.com (Scott Brim)
Date: Tue, 3 Jan 2017 11:38:48 -0500
Subject: [ih] Fwd: Reporter query on the history of greater-than in
	quoted replies
In-Reply-To: <cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
Message-ID: <CAPv4CP-drVNDHcbG5kkOMKuEv0ZbBJ6Ti8KutFbfF1uwrmM-xA@mail.gmail.com>

How were tabs shown (if shown) on early terminals?

On Tue, Jan 3, 2017 at 10:57 AM, Dave Crocker <dhc2 at dcrocker.net> wrote:
> Hi folks.
>
> Just got this query.  I don't have any clear memory of how the usage got
> started.  Do any of you?
>
> d/
>
> -------- Forwarded Message --------
> Subject: Reporter query on the history of greater-than in quoted replies
> Date: Mon,  2 Jan 2017 12:43:04 -0800
> From: Glenn Fleishman <glenn at glennf.com>
>
> ...
> I'm a reporter with a deep background in typography, and recently wrote
> this piece about quotation marks (the curly kind) for The Atlantic:
> https://www.theatlantic.com/technology/archive/2016/12/quotation-mark-wars/511766/
>
> In researching that story, I found that the > mark was the original
> quotation mark ? it dates back about two thousand years in that usage.
> This fascinated me, because of how it was revived in email and forums.
>
> I read early RFCs and other documents, and was unable to find the origin
> of > for quoting a reply.
> ...
>
>
> --
>
>    Dave Crocker
>    Brandenburg InternetWorking
>    bbiw.net
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.



From jnc at mercury.lcs.mit.edu  Tue Jan  3 09:25:02 2017
From: jnc at mercury.lcs.mit.edu (Noel Chiappa)
Date: Tue,  3 Jan 2017 12:25:02 -0500 (EST)
Subject: [ih] Reporter query on the history of greater-than in quoted
	replies
Message-ID: <20170103172502.93A9D18C0C5@mercury.lcs.mit.edu>

    >> From: Glenn Fleishman

    >> I read early RFCs and other documents, and was unable to find the
    >> origin of [']>['] for quoting a reply.

I'm pretty sure that there was no formal discussion of it, at least I don't
recall seeing any (although email wasn't my area of specialization); I
strongly suspect that someone just started doing it, and it was obviously such
a good idea, everyone copied it. (That's certainly how I started with it.)

One avenue to look at is to see if some mail reader (e.g. RMail, in Emacs)
was coded to do that on replies, and that's where people got it from.

	Noel


From bill.n1vux at gmail.com  Tue Jan  3 09:25:28 2017
From: bill.n1vux at gmail.com (Bill Ricker)
Date: Tue, 3 Jan 2017 12:25:28 -0500
Subject: [ih] Fwd: Reporter query on the history of greater-than in
	quoted replies
In-Reply-To: <CAPv4CP-drVNDHcbG5kkOMKuEv0ZbBJ6Ti8KutFbfF1uwrmM-xA@mail.gmail.com>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
	<CAPv4CP-drVNDHcbG5kkOMKuEv0ZbBJ6Ti8KutFbfF1uwrmM-xA@mail.gmail.com>
Message-ID: <CAAbKA3VHzRgm_whHWNHjCsBtWtd+U_-pMTyxWyc_sVwwAe3H+w@mail.gmail.com>

On Tue, Jan 3, 2017 at 11:38 AM, Scott Brim <scott.brim at gmail.com> wrote:

> How were tabs shown (if shown) on early terminals?


?I vaguely recall some terminals, maybe WECO TTYs?, using arrows (single
char wide) -> and <- ? for TAB and BS?
Are you suggesting back-formation from that to >  is plausible ?

( FWIW Glenn F is a good guy, and as tech savvy as tech reporter should be.
I believe he does his own programming on his hobby websites. )



-- 
Bill Ricker
bill.n1vux at gmail.com
https://www.linkedin.com/in/n1vux
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170103/3c7f1780/attachment.htm>

From brian.e.carpenter at gmail.com  Tue Jan  3 11:58:10 2017
From: brian.e.carpenter at gmail.com (Brian E Carpenter)
Date: Wed, 4 Jan 2017 08:58:10 +1300
Subject: [ih] Fwd: Reporter query on the history of greater-than in
 quoted replies
In-Reply-To: <CAAbKA3VHzRgm_whHWNHjCsBtWtd+U_-pMTyxWyc_sVwwAe3H+w@mail.gmail.com>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
	<CAPv4CP-drVNDHcbG5kkOMKuEv0ZbBJ6Ti8KutFbfF1uwrmM-xA@mail.gmail.com>
	<CAAbKA3VHzRgm_whHWNHjCsBtWtd+U_-pMTyxWyc_sVwwAe3H+w@mail.gmail.com>
Message-ID: <cbfa3fc4-f278-ab70-59f1-a0cc1bbec07b@gmail.com>

On 04/01/2017 06:25, Bill Ricker wrote:
> On Tue, Jan 3, 2017 at 11:38 AM, Scott Brim <scott.brim at gmail.com> wrote:
> 
>> How were tabs shown (if shown) on early terminals?

Surely at the relevant time, the Ur-terminal was still the ASR33? The tab
was still explicitly "move to the next tab stop" (which could be anywhere).

There was no separate tab key on the ASR33, iirc. You could use CTRL/I.

On the VT05 there was a tab key, and there were preset tab stops every 8 places.

I don't recall any convention of using > to represent tab.

    Brian

> 
> ?I vaguely recall some terminals, maybe WECO TTYs?, using arrows (single
> char wide) -> and <- ? for TAB and BS?
> Are you suggesting back-formation from that to >  is plausible ?
> 
> ( FWIW Glenn F is a good guy, and as tech savvy as tech reporter should be.
> I believe he does his own programming on his hobby websites. )
> 
> 
> 
> 
> 
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
> 




From dhc2 at dcrocker.net  Tue Jan  3 12:00:29 2017
From: dhc2 at dcrocker.net (Dave Crocker)
Date: Tue, 3 Jan 2017 12:00:29 -0800
Subject: [ih] Fwd: Reporter query on the history of greater-than in
 quoted replies
In-Reply-To: <cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
Message-ID: <89828649-f2ad-295f-f8c6-1f67c19904b3@dcrocker.net>

On 1/3/2017 7:57 AM, Dave Crocker wrote:
> Just got this query.  I don't have any clear memory of how the usage got
> started.  Do any of you?
...
> Subject: Reporter query on the history of greater-than in quoted replies
> Date: Mon,  2 Jan 2017 12:43:04 -0800
> From: Glenn Fleishman <glenn at glennf.com>
...
> I read early RFCs and other documents, and was unable to find the origin
> of > for quoting a reply.


Just looked at the 1977 documentation for the Rand MS system.  It's 
examples do not show quoted text being included.  Had the convention 
developed by then, I think it would have.

So I'm going to guess that it came up in the 80s.  Anyone have old 
copies of MH source code that might show it?

d/
-- 

   Dave Crocker
   Brandenburg InternetWorking
   bbiw.net


From mbgreen at seas.upenn.edu  Tue Jan  3 13:20:16 2017
From: mbgreen at seas.upenn.edu (Michael Greenwald)
Date: Tue, 03 Jan 2017 13:20:16 -0800
Subject: [ih] Fwd: Reporter query on the history of greater-than in
 quoted replies
In-Reply-To: <89828649-f2ad-295f-f8c6-1f67c19904b3@dcrocker.net>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
	<89828649-f2ad-295f-f8c6-1f67c19904b3@dcrocker.net>
Message-ID: <3be667aa1bce111cfeb29d1331866a98@seas.upenn.edu>

On 2017-01-03 12:00, Dave Crocker wrote:
> On 1/3/2017 7:57 AM, Dave Crocker wrote:
>> Just got this query.  I don't have any clear memory of how the usage 
>> got
>> started.  Do any of you?
> ...
>> Subject: Reporter query on the history of greater-than in quoted 
>> replies
>> Date: Mon,  2 Jan 2017 12:43:04 -0800
>> From: Glenn Fleishman <glenn at glennf.com>
> ...
>> I read early RFCs and other documents, and was unable to find the 
>> origin
>> of > for quoting a reply.
> 
> 
> Just looked at the 1977 documentation for the Rand MS system.  It's
> examples do not show quoted text being included.  Had the convention
> developed by then, I think it would have.
> 
> So I'm going to guess that it came up in the 80s.  Anyone have old
> copies of MH source code that might show it?
> 
> d/

For what it's worth, I remember that the first time I saw it (well,
the first time I *noticed* it) was after I had seen tab stops in
either TeX or Scribe, in which @> or \> was a tab stop. Second, I
think I saw it in a bulletin board before I ever saw it in email,
and remember thinking that >>, >>> etc. were better for deep
nesting than spaces.

So, it may make sense to look for sources of BBS systems, which may
predate the appearance of ">" in MH (or other electronic mail) software.

(Without concrete evidence of source file listings, of course, my
recollections are as suspect as anyone's, and should be taken with
a very large grain of salt)


From leo at vegoda.org  Tue Jan  3 13:28:05 2017
From: leo at vegoda.org (Leo Vegoda)
Date: Tue, 03 Jan 2017 13:28:05 -0800
Subject: [ih] Fwd: Reporter query on the history of greater-than in
	quoted replies
In-Reply-To: <89828649-f2ad-295f-f8c6-1f67c19904b3@dcrocker.net>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
	<89828649-f2ad-295f-f8c6-1f67c19904b3@dcrocker.net>
Message-ID: <159663b6288.27f9.d40badc7ad02dd75e2a8e01473a2ae0e@vegoda.org>

On January 3, 2017 12:07:10 PM Dave Crocker <dhc2 at dcrocker.net> wrote:

[...]

> So I'm going to guess that it came up in the 80s.  Anyone have old
> copies of MH source code that might show

I didn't use MH until 2000 but when I did I think it used * to show quoting 
by default. Perhaps that was just a site-wide configuration default, though.

Regards,

Leo




From dot at dotat.at  Wed Jan  4 05:05:55 2017
From: dot at dotat.at (Tony Finch)
Date: Wed, 4 Jan 2017 13:05:55 +0000
Subject: [ih] Fwd: Reporter query on the history of greater-than in
 quoted replies
In-Reply-To: <cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
Message-ID: <alpine.DEB.2.11.1701041301060.7102@grey.csi.cam.ac.uk>

I had a dig through the history of the BSD mail command to see if/when
message quoting turned up.

It seems to arrive in 1988, with a default indent of \t
https://github.com/dspinellis/unix-history-repo/commit/b67d96f80401cad5410d9a4cc947a14b9f4b9b12#diff-df5903b671fa9aa2e169be159339a0fa

There's documentation in this commit which suggests setting tabstr to '-> '
https://github.com/dspinellis/unix-history-repo/commit/7eec6b26cca8e69dc514d64c65385b27caec38be#diff-df5903b671fa9aa2e169be159339a0fa

In 1989 the tabstr setting was renamed to indentprefix
https://github.com/dspinellis/unix-history-repo/commit/1bc3363419be5e5ded196102481584a2931612eb#diff-df5903b671fa9aa2e169be159339a0fa
https://github.com/dspinellis/unix-history-repo/commit/3db0dac41be0491eec6c8c5b7d3d93fb99abea47#diff-df5903b671fa9aa2e169be159339a0fa

Tony.
-- 
f.anthony.n.finch  <dot at dotat.at>  http://dotat.at/  -  I xn--zr8h punycode
Trafalgar: Southerly 4 or 5, but variable 3 or less in far southeast. Moderate
or rough, occasionally slight later in far southeast. Showers. Good.


From dot at dotat.at  Wed Jan  4 05:25:43 2017
From: dot at dotat.at (Tony Finch)
Date: Wed, 4 Jan 2017 13:25:43 +0000
Subject: [ih] Fwd: Reporter query on the history of greater-than in
 quoted replies
In-Reply-To: <89828649-f2ad-295f-f8c6-1f67c19904b3@dcrocker.net>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
	<89828649-f2ad-295f-f8c6-1f67c19904b3@dcrocker.net>
Message-ID: <alpine.DEB.2.11.1701041321510.7102@grey.csi.cam.ac.uk>

Dave Crocker <dhc2 at dcrocker.net> wrote:
>
> So I'm going to guess that it came up in the 80s.  Anyone have old
> copies of MH source code that might show it?

This turned out to be a much better idea than looking at mail(1) :-)

4.3BSD isn't amazingly early, but its version of MH includes some lovely
bits about quoted replies. The READ-ME is a message dated 30 Jan 1986 from
Bob Desinger who describes it as

> prepended with those nice "> " netnews-style borders

https://github.com/dspinellis/unix-history-repo/tree/BSD-4_3-Snapshot-Development/usr/contrib/mh/miscellany/replies

Tony.
-- 
f.anthony.n.finch  <dot at dotat.at>  http://dotat.at/  -  I xn--zr8h punycode
Fitzroy, Sole: Southeasterly 4 or 5, but 6 or 7 in west. Slight or moderate,
occasionally rough in west. Occasional rain. Good, occasionally poor.


From jaap at NLnetLabs.nl  Wed Jan  4 06:46:16 2017
From: jaap at NLnetLabs.nl (Jaap Akkerhuis)
Date: Wed, 04 Jan 2017 15:46:16 +0100
Subject: [ih] Fwd: Reporter query on the history of greater-than in
	quoted replies
In-Reply-To: <alpine.DEB.2.11.1701041321510.7102@grey.csi.cam.ac.uk>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
	<89828649-f2ad-295f-f8c6-1f67c19904b3@dcrocker.net>
	<alpine.DEB.2.11.1701041321510.7102@grey.csi.cam.ac.uk>
Message-ID: <201701041446.v04EkGJn006992@bela.nlnetlabs.nl>

 Tony Finch writes:

 > 
 > 4.3BSD isn't amazingly early, but its version of MH includes some lovely
 > bits about quoted replies. The READ-ME is a message dated 30 Jan 1986 from
 > Bob Desinger who describes it as
 > 
 > > prepended with those nice "> " netnews-style borders
 > 
 > https://github.com/dspinellis/unix-history-repo/tree/BSD-4_3-Snapshot-Development/usr/contrib/mh/miscellany/replies


Yes, I do vaguely remember that the habit started with replies in
readnews from B-news but I don't have sources (nor time) to check that.

	jaap



From dhc2 at dcrocker.net  Wed Jan  4 06:46:33 2017
From: dhc2 at dcrocker.net (Dave Crocker)
Date: Wed, 4 Jan 2017 06:46:33 -0800
Subject: [ih] Fwd: Reporter query on the history of greater-than in
 quoted replies
In-Reply-To: <alpine.DEB.2.11.1701041321510.7102@grey.csi.cam.ac.uk>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
	<89828649-f2ad-295f-f8c6-1f67c19904b3@dcrocker.net>
	<alpine.DEB.2.11.1701041321510.7102@grey.csi.cam.ac.uk>
Message-ID: <4aea7e9c-86fa-2ea1-3c41-42b92e241249@dcrocker.net>

On 1/4/2017 5:25 AM, Tony Finch wrote:
>> prepended with those nice "> " netnews-style borders


Cool.

So now the question is whether there is a useful archive of netnews 
software that might allow tracking this back farther?

d/

-- 

   Dave Crocker
   Brandenburg InternetWorking
   bbiw.net


From dot at dotat.at  Wed Jan  4 07:07:38 2017
From: dot at dotat.at (Tony Finch)
Date: Wed, 4 Jan 2017 15:07:38 +0000
Subject: [ih] Fwd: Reporter query on the history of greater-than in
 quoted replies
In-Reply-To: <alpine.DEB.2.11.1701041321510.7102@grey.csi.cam.ac.uk>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
	<89828649-f2ad-295f-f8c6-1f67c19904b3@dcrocker.net>
	<alpine.DEB.2.11.1701041321510.7102@grey.csi.cam.ac.uk>
Message-ID: <alpine.DEB.2.11.1701041425380.6937@grey.csi.cam.ac.uk>

Sorry, this has turned into more of a running commentary than I expected,
but I think I have run out of places I can look now.

4.3BSD includes a couple of newsreaders, both of which have > quote
support. In rn,

https://github.com/dspinellis/unix-history-repo/blob/BSD-4_3-Snapshot-Development/usr/contrib/rn/common.h#L675
https://github.com/dspinellis/unix-history-repo/blob/BSD-4_3-Snapshot-Development/usr/contrib/rn/respond.c#L340

common.h:EXT char *indstr INIT(">");    /* indent for old article embedded in followup */

The reply() and followup() functions both have copies of the this code,

respond.c-      fseek(artfp,(long)htype[PAST_HEADER].ht_minpos,0);
respond.c-      while (fgets(buf,LBUFLEN,artfp) != Nullch) {
respond.c:          fprintf(tmpfp,"%s%s",indstr,buf);
respond.c-      }
respond.c-      fprintf(tmpfp,"\n");

The code in /usr/contrib/news/ is missing from the unix-history-repo but
it is present in Mckusick's CSRG archives. In news/src/postnews.c there is
prep_article() which does

                fprintf(tf, "In article %s, %s writes:\n", msgid, isfrom);
                while (fgets(buf, BUFSIZ, of) != NULL)
                        fprintf(tf, "> %s", buf);
                (void) fclose(of);

There is also a copy of mh-e (the Emacs interface to MH) which includes

(defvar mh-ins-buf-prefix ">> "
  "*String to put before each non-blank line of the the current message
as it is inserted in an outgoing letter.")

In the CSRG archives copy of 4.2BSD, there is no emacs nor rn. None of
mh nor the news code nor rnews.ml (emacs news reader) seem to have any
support for quoting the original message in a reply.

So it seems this convention emerged on Usenet some time between 4.2BSD
(1983) and 4.3BSD (1986).

Tony.
-- 
f.anthony.n.finch  <dot at dotat.at>  http://dotat.at/  -  I xn--zr8h punycode
Bailey: Southeast 4 or 5, veering south 6 or 7, occasionally gale 8 later.
Moderate or rough, becoming rough or very rough later. Occasional rain. Good,
occasionally poor.


From dhc2 at dcrocker.net  Wed Jan  4 07:40:32 2017
From: dhc2 at dcrocker.net (Dave Crocker)
Date: Wed, 4 Jan 2017 07:40:32 -0800
Subject: [ih] Fwd: Reporter query on the history of greater-than in
 quoted replies
In-Reply-To: <alpine.DEB.2.11.1701041425380.6937@grey.csi.cam.ac.uk>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
	<89828649-f2ad-295f-f8c6-1f67c19904b3@dcrocker.net>
	<alpine.DEB.2.11.1701041321510.7102@grey.csi.cam.ac.uk>
	<alpine.DEB.2.11.1701041425380.6937@grey.csi.cam.ac.uk>
Message-ID: <0b3db4fe-6951-8f8f-f9b5-0a754a8cac29@dcrocker.net>

On 1/4/2017 7:07 AM, Tony Finch wrote:
> Sorry, this has turned into more of a running commentary than I expected,


but excellent stuff.  thanks!

d/

-- 

   Dave Crocker
   Brandenburg InternetWorking
   bbiw.net


From scott.brim at gmail.com  Wed Jan  4 08:12:02 2017
From: scott.brim at gmail.com (Scott Brim)
Date: Wed, 4 Jan 2017 11:12:02 -0500
Subject: [ih] Fwd: Reporter query on the history of greater-than in
	quoted replies
In-Reply-To: <alpine.DEB.2.11.1701041425380.6937@grey.csi.cam.ac.uk>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
	<89828649-f2ad-295f-f8c6-1f67c19904b3@dcrocker.net>
	<alpine.DEB.2.11.1701041321510.7102@grey.csi.cam.ac.uk>
	<alpine.DEB.2.11.1701041425380.6937@grey.csi.cam.ac.uk>
Message-ID: <CAPv4CP_riH6u8bBd9g5K+3x6dY+OC4VpS_OBR4t7o2MzGmpyaA@mail.gmail.com>

I see that the disagreement about whether to put a space after ">" was
already going on. :-)

Scott

On Wed, Jan 4, 2017 at 10:07 AM, Tony Finch <dot at dotat.at> wrote:
> 4.3BSD includes a couple of newsreaders, both of which have > quote
> support. In rn,
>
> https://github.com/dspinellis/unix-history-repo/blob/BSD-4_3-Snapshot-Development/usr/contrib/rn/common.h#L675
> https://github.com/dspinellis/unix-history-repo/blob/BSD-4_3-Snapshot-Development/usr/contrib/rn/respond.c#L340
>
> common.h:EXT char *indstr INIT(">");    /* indent for old article embedded in followup */
>
> The reply() and followup() functions both have copies of the this code,
>
> respond.c-      fseek(artfp,(long)htype[PAST_HEADER].ht_minpos,0);
> respond.c-      while (fgets(buf,LBUFLEN,artfp) != Nullch) {
> respond.c:          fprintf(tmpfp,"%s%s",indstr,buf);
> respond.c-      }
> respond.c-      fprintf(tmpfp,"\n");
>
> The code in /usr/contrib/news/ is missing from the unix-history-repo but
> it is present in Mckusick's CSRG archives. In news/src/postnews.c there is
> prep_article() which does
>
>                 fprintf(tf, "In article %s, %s writes:\n", msgid, isfrom);
>                 while (fgets(buf, BUFSIZ, of) != NULL)
>                         fprintf(tf, "> %s", buf);
>                 (void) fclose(of);


From jnc at mercury.lcs.mit.edu  Wed Jan  4 08:15:38 2017
From: jnc at mercury.lcs.mit.edu (Noel Chiappa)
Date: Wed,  4 Jan 2017 11:15:38 -0500 (EST)
Subject: [ih] Reporter query on the history of greater-than in quoted
	replies
Message-ID: <20170104161538.EB55D18C0CE@mercury.lcs.mit.edu>

    > From: Tony Finch 

    > So it seems this convention emerged on Usenet some time between 4.2BSD
    > (1983) and 4.3BSD (1986).

Inclusion in a news reader is probably what made it _popular_. But there may
have been an _individual_ who came up with it, and it got taken up into the
news reader(s) after people saw it and started copying it.

So probably the place to check now is to go into Usenet archives (assuming
they are still online, I lost track of their comings and goings) and look for
instances, trying to track the behaviour backwards in time...

      Noel


From casner at acm.org  Wed Jan  4 12:55:43 2017
From: casner at acm.org (Stephen Casner)
Date: Wed, 4 Jan 2017 12:55:43 -0800 (PST)
Subject: [ih] Fwd: Reporter query on the history of greater-than in
 quoted replies
In-Reply-To: <CAPv4CP_riH6u8bBd9g5K+3x6dY+OC4VpS_OBR4t7o2MzGmpyaA@mail.gmail.com>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
	<89828649-f2ad-295f-f8c6-1f67c19904b3@dcrocker.net>
	<alpine.DEB.2.11.1701041321510.7102@grey.csi.cam.ac.uk>
	<alpine.DEB.2.11.1701041425380.6937@grey.csi.cam.ac.uk>
	<CAPv4CP_riH6u8bBd9g5K+3x6dY+OC4VpS_OBR4t7o2MzGmpyaA@mail.gmail.com>
Message-ID: <alpine.OSX.2.20.1701041252560.85837@auge.attlocal.net>

I looked for examples of the angle bracket used for quoting in my
personal email archive.  It only contains selected messages from the
early days since disk space was still considered expensive then.

The oldest example I found was from January, 1985 that showed two
levels of quoting where the whole message body was included at both
levels and the first line was a ">From:" line.  The inner quote (from
rochester.arpa) included a space after the bracket, while the outer
one (from bbn-vax.arpa) did not.

Another example from April 1985 between two of my colleagues at ISI
includes quoting of the Date, Subject, From and To header lines in
addition to the message body.  This would have been on TOPS-20.  In
those days the email program did not announce itself by adding a
header line of its own, but I observe that the outer message contains
an "In-Reply-To" header as well as non-standard Phone, Address and
Never headers (as in "Never: Play pool with a man named Fats").  I
don't know if those hints help identify the program, but I remember
using MM in that era.  There was some local development on MM at ISI.

For both of these examples the quoting was most likely done by the
email program, but I would guess that the practice started with manual
editing before the programs implemented the feature.  In an example
message from March 1985 I saw angle brackets used to offset a couple
of comments added by the sender as editor in a message that collected
a thread of messages.  In that case it was anti-quoting.

I also observe that in both of these examples the new text was added
below the quoted message rather than above.

One other precursor that I observed was the following example line
from 1981:

    >From decvax!duke!unc!smb  Fri Jun  5 08:45:48 1981

This looks like an example of the MTA inserting the bracket to avoid
having the word "From" at the beginning of a line be detected as the
envelope field marking the start of a message in the email file.  That
practice continues today.  Perhaps this use of the bracket came first
and suggested the use for quoting of nested messages.

                                                        -- Steve


From casner at acm.org  Wed Jan  4 13:08:33 2017
From: casner at acm.org (Stephen Casner)
Date: Wed, 4 Jan 2017 13:08:33 -0800 (PST)
Subject: [ih] Fwd: Reporter query on the history of greater-than in
 quoted replies
In-Reply-To: <alpine.OSX.2.20.1701041252560.85837@auge.attlocal.net>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
	<89828649-f2ad-295f-f8c6-1f67c19904b3@dcrocker.net>
	<alpine.DEB.2.11.1701041321510.7102@grey.csi.cam.ac.uk>
	<alpine.DEB.2.11.1701041425380.6937@grey.csi.cam.ac.uk>
	<CAPv4CP_riH6u8bBd9g5K+3x6dY+OC4VpS_OBR4t7o2MzGmpyaA@mail.gmail.com>
	<alpine.OSX.2.20.1701041252560.85837@auge.attlocal.net>
Message-ID: <alpine.OSX.2.20.1701041303450.85837@auge.attlocal.net>

On Wed, 4 Jan 2017, Stephen Casner wrote:

> One other precursor that I observed was the following example line
> from 1981:
>
>     >From decvax!duke!unc!smb  Fri Jun  5 08:45:48 1981
>
> This looks like an example of the MTA inserting the bracket to avoid
> having the word "From" at the beginning of a line be detected as the
> envelope field marking the start of a message in the email file.  That
> practice continues today.  Perhaps this use of the bracket came first
> and suggested the use for quoting of nested messages.

I should add that this was a case where a message was inserted
wholesale into a reply without any quoting, hence the inclusion of the
envelope From line.

The other examples that I found having angle brackets at the left
margin in email before 1985 were cases where a portion of a
command-line typescript was included in the message and the prompt
character for the CLI was the angle bracket.  That may have been
another source contributing to the uptake for quoting purposes.

                                                        -- Steve


From brian.e.carpenter at gmail.com  Wed Jan  4 16:25:13 2017
From: brian.e.carpenter at gmail.com (Brian E Carpenter)
Date: Thu, 5 Jan 2017 13:25:13 +1300
Subject: [ih] Fwd: Reporter query on the history of greater-than in
 quoted replies
In-Reply-To: <alpine.OSX.2.20.1701041303450.85837@auge.attlocal.net>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
	<89828649-f2ad-295f-f8c6-1f67c19904b3@dcrocker.net>
	<alpine.DEB.2.11.1701041321510.7102@grey.csi.cam.ac.uk>
	<alpine.DEB.2.11.1701041425380.6937@grey.csi.cam.ac.uk>
	<CAPv4CP_riH6u8bBd9g5K+3x6dY+OC4VpS_OBR4t7o2MzGmpyaA@mail.gmail.com>
	<alpine.OSX.2.20.1701041252560.85837@auge.attlocal.net>
	<alpine.OSX.2.20.1701041303450.85837@auge.attlocal.net>
Message-ID: <ee77263a-3c3e-5c7b-c946-f074bc859ea8@gmail.com>

This page suggests to me that finding the earliest documentation of the Unix mbox format would be instructive:

https://www.freebsd.org/cgi/man.cgi?query=mbox&apropos=0&sektion=0&manpath=FreeBSD+Ports&format=html

Regards
   Brian


From dot at dotat.at  Wed Jan  4 16:50:08 2017
From: dot at dotat.at (Tony Finch)
Date: Thu, 5 Jan 2017 00:50:08 +0000
Subject: [ih] Fwd: Reporter query on the history of greater-than in
	quoted replies
In-Reply-To: <ee77263a-3c3e-5c7b-c946-f074bc859ea8@gmail.com>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
	<89828649-f2ad-295f-f8c6-1f67c19904b3@dcrocker.net>
	<alpine.DEB.2.11.1701041321510.7102@grey.csi.cam.ac.uk>
	<alpine.DEB.2.11.1701041425380.6937@grey.csi.cam.ac.uk>
	<CAPv4CP_riH6u8bBd9g5K+3x6dY+OC4VpS_OBR4t7o2MzGmpyaA@mail.gmail.com>
	<alpine.OSX.2.20.1701041252560.85837@auge.attlocal.net>
	<alpine.OSX.2.20.1701041303450.85837@auge.attlocal.net>
	<ee77263a-3c3e-5c7b-c946-f074bc859ea8@gmail.com>
Message-ID: <E1564C5F-AA26-49B3-ACA8-24A59E0C0B71@dotat.at>


> On 5 Jan 2017, at 00:25, Brian E Carpenter <brian.e.carpenter at gmail.com> wrote:
> 
> This page suggests to me that finding the earliest documentation of the Unix mbox format would be instructive:

The mbox format is mostly ignorant of how message bodies are formatted: it doesn't care about reply quoting conventions or MIME or whatever. But the ugly mbox >From hack does go back a long way, before 4.2BSD. I have no idea if it had any influence on how users decided to format replies...

BTW, If anyone else wants to dig through old code, I found it surprisingly effective to simply

        grep -R '">' .

i.e. look for string constants starting with the quote marker. A lot of the false positives were the >From hack, or interactive prompts.

Tony.
-- 
f.anthony.n.finch  <dot at dotat.at>  http://dotat.at





From demco at webnames.ca  Wed Jan  4 17:17:12 2017
From: demco at webnames.ca (John Demco)
Date: Wed, 4 Jan 2017 17:17:12 -0800
Subject: [ih] Fwd: Reporter query on the history of greater-than in
 quoted replies
In-Reply-To: <E1564C5F-AA26-49B3-ACA8-24A59E0C0B71@dotat.at>
References: <r473Ps-10122i-AA8AAD87C0844CD2A898D846CBDB57B9@Mininum-Minival.local>
	<cb3a8aef-1dec-80aa-32bb-f4cd9c71746e@dcrocker.net>
	<89828649-f2ad-295f-f8c6-1f67c19904b3@dcrocker.net>
	<alpine.DEB.2.11.1701041321510.7102@grey.csi.cam.ac.uk>
	<alpine.DEB.2.11.1701041425380.6937@grey.csi.cam.ac.uk>
	<CAPv4CP_riH6u8bBd9g5K+3x6dY+OC4VpS_OBR4t7o2MzGmpyaA@mail.gmail.com>
	<alpine.OSX.2.20.1701041252560.85837@auge.attlocal.net>
	<alpine.OSX.2.20.1701041303450.85837@auge.attlocal.net>
	<ee77263a-3c3e-5c7b-c946-f074bc859ea8@gmail.com>
	<E1564C5F-AA26-49B3-ACA8-24A59E0C0B71@dotat.at>
Message-ID: <3d8f1dd1-387b-37d3-fee8-b1152af2808e@webnames.ca>

An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170104/d6ce96fa/attachment.htm>

From pnr at planet.nl  Wed Jan 18 01:10:21 2017
From: pnr at planet.nl (Paul Ruizendaal)
Date: Wed, 18 Jan 2017 10:10:21 +0100
Subject: [ih] Arpanet line speed
Message-ID: <F559B87A-BAD5-4403-97CD-75DA75F7DF84@planet.nl>


I'm wondering about (early) Arpanet line speed. Online sources seem to say that it was 50kb/s on the long haul connections, provided by AT&T. To me, that "50" sounds like a strange number:

- if the modem from the IMP to the Bell System was analog, the best technology of the time was perhaps 2.4kb/s (and this speed seems to have been considered initially). In this case a speed of 50kb/s is not a multiple of 2.4kb/s, and it would have required 21 parallel lines -- more logical would have been to use a full T1 connection and have 24 parallel lines or 57.6kb/s.

- if the modem from the IMP to the Bell System was digital, it would most likely have used a single channel of a T1 connection, i.e. a 56kb/s rate. I suppose AT&T would have to run special lines from the IMP to the trunk, as I don't think digital lines were used for the last mile back then (perhaps they were to company/hotel PBX's, I don't know).

Perhaps the "50" is a rounded number, perhaps some capacity is used for error correction, perhaps it is plain wrong.

Does anybody have more background information on this?

Thanks,

Paul






From vint at google.com  Wed Jan 18 02:23:20 2017
From: vint at google.com (Vint Cerf)
Date: Wed, 18 Jan 2017 05:23:20 -0500
Subject: [ih] Arpanet line speed
In-Reply-To: <F559B87A-BAD5-4403-97CD-75DA75F7DF84@planet.nl>
References: <F559B87A-BAD5-4403-97CD-75DA75F7DF84@planet.nl>
Message-ID: <CAHxHggcXmiNzi-Au2Loo0usXGiSmY8b9ukSp7MaxgDdBKGrHAw@mail.gmail.com>

no it was really 50 kb/s. Used Bell 303A modems as I recall. A portion of
the BW was for a supervisory voice channel I think (?).  12 voice channels
were bonded producing 36 Khz of analog bandwidth.

Bob Kahn will probably have a more comprehensive answer.

v


On Wed, Jan 18, 2017 at 4:10 AM, Paul Ruizendaal <pnr at planet.nl> wrote:

>
> I'm wondering about (early) Arpanet line speed. Online sources seem to say
> that it was 50kb/s on the long haul connections, provided by AT&T. To me,
> that "50" sounds like a strange number:
>
> - if the modem from the IMP to the Bell System was analog, the best
> technology of the time was perhaps 2.4kb/s (and this speed seems to have
> been considered initially). In this case a speed of 50kb/s is not a
> multiple of 2.4kb/s, and it would have required 21 parallel lines -- more
> logical would have been to use a full T1 connection and have 24 parallel
> lines or 57.6kb/s.
>
> - if the modem from the IMP to the Bell System was digital, it would most
> likely have used a single channel of a T1 connection, i.e. a 56kb/s rate. I
> suppose AT&T would have to run special lines from the IMP to the trunk, as
> I don't think digital lines were used for the last mile back then (perhaps
> they were to company/hotel PBX's, I don't know).
>
> Perhaps the "50" is a rounded number, perhaps some capacity is used for
> error correction, perhaps it is plain wrong.
>
> Does anybody have more background information on this?
>
> Thanks,
>
> Paul
>
>
>
>
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
>



-- 
New postal address:
Google
1875 Explorer Street, 10th Floor
Reston, VA 20190
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170118/61c8277b/attachment.htm>

From ats at offog.org  Wed Jan 18 03:10:46 2017
From: ats at offog.org (Adam Sampson)
Date: Wed, 18 Jan 2017 11:10:46 +0000
Subject: [ih] Arpanet line speed
In-Reply-To: <F559B87A-BAD5-4403-97CD-75DA75F7DF84@planet.nl> (Paul
	Ruizendaal's message of "Wed, 18 Jan 2017 10:10:21 +0100")
References: <F559B87A-BAD5-4403-97CD-75DA75F7DF84@planet.nl>
Message-ID: <y2a1sw0y66h.fsf@offog.org>

Paul Ruizendaal <pnr at planet.nl> writes:

> I'm wondering about (early) Arpanet line speed. Online sources seem to
> say that it was 50kb/s on the long haul connections, provided by
> AT&T. To me, that "50" sounds like a strange number:

This has been discussed previously on this list:
http://mailman.postel.org/pipermail/internet-history/2004-February/000384.html

The August 1966 manual for the Bell 303 Wideband Data Station used to
implement these links is available from bitsavers, and gives a pretty
good description of how it works:
http://bitsavers.org/pdf/westernElectric/telephony/modems/

There are a number of mentions in there of "50 kbit per second operation
with a balanced interface for use with government cryptographic
equipment", which I assume explains the odd speed (but it really does
have a 50 kHz synchronous clock rate!). The other speeds it supports
look more like "round" numbers to the modern eye.

Thanks,

-- 
Adam Sampson <ats at offog.org>                         <http://offog.org/>


From jeanjour at comcast.net  Wed Jan 18 03:11:30 2017
From: jeanjour at comcast.net (John Day)
Date: Wed, 18 Jan 2017 06:11:30 -0500
Subject: [ih] Arpanet line speed
In-Reply-To: <F559B87A-BAD5-4403-97CD-75DA75F7DF84@planet.nl>
References: <F559B87A-BAD5-4403-97CD-75DA75F7DF84@planet.nl>
Message-ID: <A65B141F-FA38-464B-9961-CE6CBDC96471@comcast.net>

I believe it was 56Kb. A DS0 out of a T1.

John

> On Jan 18, 2017, at 04:10, Paul Ruizendaal <pnr at planet.nl> wrote:
> 
> 
> I'm wondering about (early) Arpanet line speed. Online sources seem to say that it was 50kb/s on the long haul connections, provided by AT&T. To me, that "50" sounds like a strange number:
> 
> - if the modem from the IMP to the Bell System was analog, the best technology of the time was perhaps 2.4kb/s (and this speed seems to have been considered initially). In this case a speed of 50kb/s is not a multiple of 2.4kb/s, and it would have required 21 parallel lines -- more logical would have been to use a full T1 connection and have 24 parallel lines or 57.6kb/s.
> 
> - if the modem from the IMP to the Bell System was digital, it would most likely have used a single channel of a T1 connection, i.e. a 56kb/s rate. I suppose AT&T would have to run special lines from the IMP to the trunk, as I don't think digital lines were used for the last mile back then (perhaps they were to company/hotel PBX's, I don't know).
> 
> Perhaps the "50" is a rounded number, perhaps some capacity is used for error correction, perhaps it is plain wrong.
> 
> Does anybody have more background information on this?
> 
> Thanks,
> 
> Paul
> 
> 
> 
> 
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.




From dave.walden.family at gmail.com  Wed Jan 18 05:06:51 2017
From: dave.walden.family at gmail.com (dave.walden.family at gmail.com)
Date: Wed, 18 Jan 2017 08:06:51 -0500
Subject: [ih] Arpanet line speed
In-Reply-To: <F559B87A-BAD5-4403-97CD-75DA75F7DF84@planet.nl>
References: <F559B87A-BAD5-4403-97CD-75DA75F7DF84@planet.nl>
Message-ID: <A206C71A-85A0-4E39-BB3E-059DF645F55F@gmail.com>

Right, Bob Kahn would know details as he dealt with ATT and specified the 24-bit CRC.  Ben Barker would also know, but he may not be on this list.  The following may be helpful:  
  http://walden-family.com/impcode/imp-hardware.pdf
pages 4-1 to 4-5. (BBN designed the IMP/MODEM interface,, Honeywell implemented it using their board technology, Ben and Severo Ornstein debugged the Honeywell implementation, and then had a test program, I believe, to check later machine deliveries for correct implementation as they arrived.)

On Jan 18, 2017, at 4:10 AM, Paul Ruizendaal <pnr at planet.nl> wrote:

> 
> I'm wondering about (early) Arpanet line speed. Online sources seem to say that it was 50kb/s on the long haul connections, provided by AT&T. To me, that "50" sounds like a strange number:
> 
> - if the modem from the IMP to the Bell System was analog, the best technology of the time was perhaps 2.4kb/s (and this speed seems to have been considered initially). In this case a speed of 50kb/s is not a multiple of 2.4kb/s, and it would have required 21 parallel lines -- more logical would have been to use a full T1 connection and have 24 parallel lines or 57.6kb/s.
> 
> - if the modem from the IMP to the Bell System was digital, it would most likely have used a single channel of a T1 connection, i.e. a 56kb/s rate. I suppose AT&T would have to run special lines from the IMP to the trunk, as I don't think digital lines were used for the last mile back then (perhaps they were to company/hotel PBX's, I don't know).
> 
> Perhaps the "50" is a rounded number, perhaps some capacity is used for error correction, perhaps it is plain wrong.
> 
> Does anybody have more background information on this?
> 
> Thanks,
> 
> Paul
> 
> 
> 
> 
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.



From pnr at planet.nl  Wed Jan 18 06:09:52 2017
From: pnr at planet.nl (Paul Ruizendaal)
Date: Wed, 18 Jan 2017 15:09:52 +0100
Subject: [ih] Arpanet line speed
In-Reply-To: <A206C71A-85A0-4E39-BB3E-059DF645F55F@gmail.com>
References: <F559B87A-BAD5-4403-97CD-75DA75F7DF84@planet.nl>
	<A206C71A-85A0-4E39-BB3E-059DF645F55F@gmail.com>
Message-ID: <ECBD0EEE-6B31-40EE-B07A-FC6937CCBAAA@planet.nl>

Many thanks, all. Most enlightening!

Paul



From jnc at mercury.lcs.mit.edu  Wed Jan 18 06:29:23 2017
From: jnc at mercury.lcs.mit.edu (Noel Chiappa)
Date: Wed, 18 Jan 2017 09:29:23 -0500 (EST)
Subject: [ih] Arpanet line speed
Message-ID: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>

    > From: Paul Ruizendaal

    > - if the modem from the IMP to the Bell System was analog, the best
    > technology of the time was perhaps 2.4kb/s ... a speed of 50kb/s is not
    > a multiple of 2.4kb/s, and it would have required 21 parallel lines

If you read the 303 manual, it's clear that i) the signal between a pair of
303 modems was analog, not digital, and ii) there was a single line, with a
wide enough bandpass to carry signals of high enough frequency to carry that
bit rate - it didn't glue together a bunch of slower lines.

    > if the modem from the IMP to the Bell System was digital, it would most
    > likely have used a single channel of a T1 connection

The whole T hierarchy was just getting started then (initial deployment in the
early 1960s), and I'm not sure if it was deployed widely enough to have made
it possible to lease a T1 line from one coast to another.

Also, many of these lines would have crossed non-AT+T local phone companies
(the Bell System did not control all of the US phone system, although some
people don't realize that). The "History of the ARPANET: The First Decade"
(which I have previously pointed you at on another list), pg. III-32, says
"In the case of a circuit from UCLA to RAND ... the service would be procured
from General Telephone" - GT was the largest independent telephone company in
the US at that point. It's not clear that those local carriers would have
supported T1.


Moral of the story: when doing history, it's bad to make assumptions about
what was and wasn't possible, and about what did and did not happen. Find
contemporary documentation.

	Noel


From amckenzie3 at yahoo.com  Wed Jan 18 06:55:23 2017
From: amckenzie3 at yahoo.com (Alex McKenzie)
Date: Wed, 18 Jan 2017 14:55:23 +0000 (UTC)
Subject: [ih] Arpanet line speed
In-Reply-To: <A65B141F-FA38-464B-9961-CE6CBDC96471@comcast.net>
References: <F559B87A-BAD5-4403-97CD-75DA75F7DF84@planet.nl>
	<A65B141F-FA38-464B-9961-CE6CBDC96471@comcast.net>
Message-ID: <1939695831.8474699.1484751323244@mail.yahoo.com>

John,
The Bell 303 modem was specified to run at 50kbps.? Your guess about DS0 sounds plausible for the underlying circuit speed but I don't know if it is correct.
Cheers,Alex


      From: John Day <jeanjour at comcast.net>
 To: Paul Ruizendaal <pnr at planet.nl> 
Cc: internet-history at postel.org
 Sent: Wednesday, January 18, 2017 6:11 AM
 Subject: Re: [ih] Arpanet line speed
   
I believe it was 56Kb. A DS0 out of a T1.

John

> On Jan 18, 2017, at 04:10, Paul Ruizendaal <pnr at planet.nl> wrote:
> 
> 
> I'm wondering about (early) Arpanet line speed. Online sources seem to say that it was 50kb/s on the long haul connections, provided by AT&T. To me, that "50" sounds like a strange number:
> 
> - if the modem from the IMP to the Bell System was analog, the best technology of the time was perhaps 2.4kb/s (and this speed seems to have been considered initially). In this case a speed of 50kb/s is not a multiple of 2.4kb/s, and it would have required 21 parallel lines -- more logical would have been to use a full T1 connection and have 24 parallel lines or 57.6kb/s.
> 
> - if the modem from the IMP to the Bell System was digital, it would most likely have used a single channel of a T1 connection, i.e. a 56kb/s rate. I suppose AT&T would have to run special lines from the IMP to the trunk, as I don't think digital lines were used for the last mile back then (perhaps they were to company/hotel PBX's, I don't know).
> 
> Perhaps the "50" is a rounded number, perhaps some capacity is used for error correction, perhaps it is plain wrong.
> 
> Does anybody have more background information on this?
> 
> Thanks,
> 
> Paul
> 
> 
> 
> 
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.


_______
internet-history mailing list
internet-history at postel.org
http://mailman.postel.org/mailman/listinfo/internet-history
Contact list-owner at postel.org for assistance.


   
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170118/229d4fcf/attachment.htm>

From vint at google.com  Wed Jan 18 06:57:40 2017
From: vint at google.com (Vint Cerf)
Date: Wed, 18 Jan 2017 09:57:40 -0500
Subject: [ih] Arpanet line speed
In-Reply-To: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
References: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
Message-ID: <CAHxHggc8g4Y8n3xcoaXOtt2+nVjaqbqpReBjtpTt4=YeN25ywA@mail.gmail.com>

Noel, wasn't it GTE (General Telephone and Electric)?

v


On Wed, Jan 18, 2017 at 9:29 AM, Noel Chiappa <jnc at mercury.lcs.mit.edu>
wrote:

>     > From: Paul Ruizendaal
>
>     > - if the modem from the IMP to the Bell System was analog, the best
>     > technology of the time was perhaps 2.4kb/s ... a speed of 50kb/s is
> not
>     > a multiple of 2.4kb/s, and it would have required 21 parallel lines
>
> If you read the 303 manual, it's clear that i) the signal between a pair of
> 303 modems was analog, not digital, and ii) there was a single line, with a
> wide enough bandpass to carry signals of high enough frequency to carry
> that
> bit rate - it didn't glue together a bunch of slower lines.
>
>     > if the modem from the IMP to the Bell System was digital, it would
> most
>     > likely have used a single channel of a T1 connection
>
> The whole T hierarchy was just getting started then (initial deployment in
> the
> early 1960s), and I'm not sure if it was deployed widely enough to have
> made
> it possible to lease a T1 line from one coast to another.
>
> Also, many of these lines would have crossed non-AT+T local phone companies
> (the Bell System did not control all of the US phone system, although some
> people don't realize that). The "History of the ARPANET: The First Decade"
> (which I have previously pointed you at on another list), pg. III-32, says
> "In the case of a circuit from UCLA to RAND ... the service would be
> procured
> from General Telephone" - GT was the largest independent telephone company
> in
> the US at that point. It's not clear that those local carriers would have
> supported T1.
>
>
> Moral of the story: when doing history, it's bad to make assumptions about
> what was and wasn't possible, and about what did and did not happen. Find
> contemporary documentation.
>
>         Noel
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
>



-- 
New postal address:
Google
1875 Explorer Street, 10th Floor
Reston, VA 20190
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170118/1964f0fc/attachment.htm>

From agmalis at gmail.com  Wed Jan 18 06:59:06 2017
From: agmalis at gmail.com (Andrew G. Malis)
Date: Wed, 18 Jan 2017 09:59:06 -0500
Subject: [ih] Arpanet line speed
In-Reply-To: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
References: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
Message-ID: <CAA=duU1u37s_umFsq3pdSsLMz+=KCEUb=+wcD_BzeOkC3Zhj=Q@mail.gmail.com>

By the time I came on board in '79, almost all of the links were 56 Kbps,
with a few 9.6 Kbps links here and there. The 50 Kbps links had been
replaced by that point.

Cheers,
Andy


On Wed, Jan 18, 2017 at 9:29 AM, Noel Chiappa <jnc at mercury.lcs.mit.edu>
wrote:

>     > From: Paul Ruizendaal
>
>     > - if the modem from the IMP to the Bell System was analog, the best
>     > technology of the time was perhaps 2.4kb/s ... a speed of 50kb/s is
> not
>     > a multiple of 2.4kb/s, and it would have required 21 parallel lines
>
> If you read the 303 manual, it's clear that i) the signal between a pair of
> 303 modems was analog, not digital, and ii) there was a single line, with a
> wide enough bandpass to carry signals of high enough frequency to carry
> that
> bit rate - it didn't glue together a bunch of slower lines.
>
>     > if the modem from the IMP to the Bell System was digital, it would
> most
>     > likely have used a single channel of a T1 connection
>
> The whole T hierarchy was just getting started then (initial deployment in
> the
> early 1960s), and I'm not sure if it was deployed widely enough to have
> made
> it possible to lease a T1 line from one coast to another.
>
> Also, many of these lines would have crossed non-AT+T local phone companies
> (the Bell System did not control all of the US phone system, although some
> people don't realize that). The "History of the ARPANET: The First Decade"
> (which I have previously pointed you at on another list), pg. III-32, says
> "In the case of a circuit from UCLA to RAND ... the service would be
> procured
> from General Telephone" - GT was the largest independent telephone company
> in
> the US at that point. It's not clear that those local carriers would have
> supported T1.
>
>
> Moral of the story: when doing history, it's bad to make assumptions about
> what was and wasn't possible, and about what did and did not happen. Find
> contemporary documentation.
>
>         Noel
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170118/6a66cd17/attachment.htm>

From craig at tereschau.net  Wed Jan 18 07:21:53 2017
From: craig at tereschau.net (Craig Partridge)
Date: Wed, 18 Jan 2017 10:21:53 -0500
Subject: [ih] Arpanet line speed
In-Reply-To: <CAA=duU1u37s_umFsq3pdSsLMz+=KCEUb=+wcD_BzeOkC3Zhj=Q@mail.gmail.com>
References: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
	<CAA=duU1u37s_umFsq3pdSsLMz+=KCEUb=+wcD_BzeOkC3Zhj=Q@mail.gmail.com>
Message-ID: <CAHQj4CdtB6LuCU5iv85DhcM=cPK=_aDVv1pSEQXXGdLGNbU86w@mail.gmail.com>

Andy's recollection that stuff was 56 Kbps later on matches my recollection
when I came on board in '83.

One thing is I don't remember what the bandwidth was on the satellite links
to England.  My recollection is it was
higher than 9.6Kbps but there was some oddity in integrating European and
US telecom standards such that the bandwidth was different.  But this could
be entirely wrong -- alas the ARPANET maps don't tell me the data rates.

Thanks!

Craig

On Wed, Jan 18, 2017 at 9:59 AM, Andrew G. Malis <agmalis at gmail.com> wrote:

> By the time I came on board in '79, almost all of the links were 56 Kbps,
> with a few 9.6 Kbps links here and there. The 50 Kbps links had been
> replaced by that point.
>
> Cheers,
> Andy
>
>
> On Wed, Jan 18, 2017 at 9:29 AM, Noel Chiappa <jnc at mercury.lcs.mit.edu>
> wrote:
>
>>     > From: Paul Ruizendaal
>>
>>     > - if the modem from the IMP to the Bell System was analog, the best
>>     > technology of the time was perhaps 2.4kb/s ... a speed of 50kb/s is
>> not
>>     > a multiple of 2.4kb/s, and it would have required 21 parallel lines
>>
>> If you read the 303 manual, it's clear that i) the signal between a pair
>> of
>> 303 modems was analog, not digital, and ii) there was a single line, with
>> a
>> wide enough bandpass to carry signals of high enough frequency to carry
>> that
>> bit rate - it didn't glue together a bunch of slower lines.
>>
>>     > if the modem from the IMP to the Bell System was digital, it would
>> most
>>     > likely have used a single channel of a T1 connection
>>
>> The whole T hierarchy was just getting started then (initial deployment
>> in the
>> early 1960s), and I'm not sure if it was deployed widely enough to have
>> made
>> it possible to lease a T1 line from one coast to another.
>>
>> Also, many of these lines would have crossed non-AT+T local phone
>> companies
>> (the Bell System did not control all of the US phone system, although some
>> people don't realize that). The "History of the ARPANET: The First Decade"
>> (which I have previously pointed you at on another list), pg. III-32, says
>> "In the case of a circuit from UCLA to RAND ... the service would be
>> procured
>> from General Telephone" - GT was the largest independent telephone
>> company in
>> the US at that point. It's not clear that those local carriers would have
>> supported T1.
>>
>>
>> Moral of the story: when doing history, it's bad to make assumptions about
>> what was and wasn't possible, and about what did and did not happen. Find
>> contemporary documentation.
>>
>>         Noel
>> _______
>> internet-history mailing list
>> internet-history at postel.org
>> http://mailman.postel.org/mailman/listinfo/internet-history
>> Contact list-owner at postel.org for assistance.
>>
>
>
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
>
>


-- 
*****
Craig Partridge's email account for professional society activities and
mailing lists.
For Raytheon business, please email: craig. <craig at bbn.com>
partridge at raytheon.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170118/d138a0c4/attachment.htm>

From jnc at mercury.lcs.mit.edu  Wed Jan 18 07:29:39 2017
From: jnc at mercury.lcs.mit.edu (Noel Chiappa)
Date: Wed, 18 Jan 2017 10:29:39 -0500 (EST)
Subject: [ih] Arpanet line speed
Message-ID: <20170118152939.C92FD18C09A@mercury.lcs.mit.edu>

    > From: Vint Cerf

    >> The "History of the ARPANET: The First Decade" ... pg. III-32, says "In
    >> the case of a circuit from UCLA to RAND ... the service would be
    >> procured from General Telephone" - GT was the largest independent
    >> telephone company in the US at that point.

    > Noel, wasn't it GTE (General Telephone and Electric)?

It seems to be complicated! The "General Telephone" is what's in the "History"
document. According to:

  https://en.wikipedia.org/wiki/GTE

(not the most reliable source, I know, but I don't have anything better on hand):

  "General Telephone's holdings included 15 telephone companies across 20
  states by 1951 ... In 1959, General Telephone and Sylvania Electric Products
  merged, and the parent's name was changed to General Telephone & Electric
  Corporation"

So as of the mid-60's, perhaps the holding company was called 'GTE', but the
local subsidiaries were still plain 'GT'?

The one in question was likely "General Telephone of California", which seemed
to still exist under that name in the late 80's:

 https://www.fcc.gov/document/application-general-telephone-california-operate-broadband-transport-facilities-cerritos-ca

but from what I can glean online in a quick look, it changed its name to "GTE
California" in 1988.  Apparently it was originally "The Associated Telephone
Company", and changed its name to General Telephone of California in 1952
(probably after being acquired by GT - as it was called then).

      Noel


From dhc2 at dcrocker.net  Wed Jan 18 07:31:26 2017
From: dhc2 at dcrocker.net (Dave Crocker)
Date: Wed, 18 Jan 2017 07:31:26 -0800
Subject: [ih] Arpanet line speed
In-Reply-To: <CAA=duU1u37s_umFsq3pdSsLMz+=KCEUb=+wcD_BzeOkC3Zhj=Q@mail.gmail.com>
References: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
	<CAA=duU1u37s_umFsq3pdSsLMz+=KCEUb=+wcD_BzeOkC3Zhj=Q@mail.gmail.com>
Message-ID: <174e102a-9948-6633-a9f6-f188daed09ad@dcrocker.net>

On 1/18/2017 6:59 AM, Andrew G. Malis wrote:
> By the time I came on board in '79, almost all of the links were 56
> Kbps, with a few 9.6 Kbps links here and there. The 50 Kbps links had
> been replaced by that point.
>


My recollection is that all the initial lines (for some years) were the 
same, 50Kbps, speed, to keep the overall system dynamics simpler, and 
that 9.6 lines were only allowed much later.  Might have been by 79, though.

d/

-- 

   Dave Crocker
   Brandenburg InternetWorking
   bbiw.net


From dave.walden.family at gmail.com  Wed Jan 18 07:39:17 2017
From: dave.walden.family at gmail.com (dave.walden.family at gmail.com)
Date: Wed, 18 Jan 2017 10:39:17 -0500
Subject: [ih] Arpanet line speed
In-Reply-To: <CAHxHggc8g4Y8n3xcoaXOtt2+nVjaqbqpReBjtpTt4=YeN25ywA@mail.gmail.com>
References: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
	<CAHxHggc8g4Y8n3xcoaXOtt2+nVjaqbqpReBjtpTt4=YeN25ywA@mail.gmail.com>
Message-ID: <880BEBAA-323D-467C-821C-1A8683F4C8CC@gmail.com>

See the text immediately below Figure 1 on the second page at 
   http://walden-family.com/public/1970-imp-afips.pdf
probably drafted by Bob Kahn.

Sent from my iPad

On Jan 18, 2017, at 9:57 AM, Vint Cerf <vint at google.com> wrote:

> Noel, wasn't it GTE (General Telephone and Electric)?
> 
> v
> 
> 
> On Wed, Jan 18, 2017 at 9:29 AM, Noel Chiappa <jnc at mercury.lcs.mit.edu> wrote:
>     > From: Paul Ruizendaal
> 
>     > - if the modem from the IMP to the Bell System was analog, the best
>     > technology of the time was perhaps 2.4kb/s ... a speed of 50kb/s is not
>     > a multiple of 2.4kb/s, and it would have required 21 parallel lines
> 
> If you read the 303 manual, it's clear that i) the signal between a pair of
> 303 modems was analog, not digital, and ii) there was a single line, with a
> wide enough bandpass to carry signals of high enough frequency to carry that
> bit rate - it didn't glue together a bunch of slower lines.
> 
>     > if the modem from the IMP to the Bell System was digital, it would most
>     > likely have used a single channel of a T1 connection
> 
> The whole T hierarchy was just getting started then (initial deployment in the
> early 1960s), and I'm not sure if it was deployed widely enough to have made
> it possible to lease a T1 line from one coast to another.
> 
> Also, many of these lines would have crossed non-AT+T local phone companies
> (the Bell System did not control all of the US phone system, although some
> people don't realize that). The "History of the ARPANET: The First Decade"
> (which I have previously pointed you at on another list), pg. III-32, says
> "In the case of a circuit from UCLA to RAND ... the service would be procured
> from General Telephone" - GT was the largest independent telephone company in
> the US at that point. It's not clear that those local carriers would have
> supported T1.
> 
> 
> Moral of the story: when doing history, it's bad to make assumptions about
> what was and wasn't possible, and about what did and did not happen. Find
> contemporary documentation.
> 
>         Noel
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
> 
> 
> 
> -- 
> New postal address:
> Google
> 1875 Explorer Street, 10th Floor
> Reston, VA 20190
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170118/a0771f85/attachment.htm>

From dave.walden.family at gmail.com  Wed Jan 18 08:02:58 2017
From: dave.walden.family at gmail.com (dave.walden.family at gmail.com)
Date: Wed, 18 Jan 2017 11:02:58 -0500
Subject: [ih] Arpanet line speed
In-Reply-To: <174e102a-9948-6633-a9f6-f188daed09ad@dcrocker.net>
References: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
	<CAA=duU1u37s_umFsq3pdSsLMz+=KCEUb=+wcD_BzeOkC3Zhj=Q@mail.gmail.com>
	<174e102a-9948-6633-a9f6-f188daed09ad@dcrocker.net>
Message-ID: <AFB93CD1-1B62-46D9-A219-2D0A0744C2F4@gmail.com>

The satellite links to Europe and Hawaii by 1973 were maybe 9.6kbs.   By that time (or 1975) the IMP code knew about four different line speeds (as I remember) -- 9.6, 50, 300 (I think) and something else (I can look it up in the code). The IMP had to know the line speeds as part of deciding if a link was up or down based on how long it was between receiving instances of interIMP traffic.




From amckenzie3 at yahoo.com  Wed Jan 18 08:03:34 2017
From: amckenzie3 at yahoo.com (Alex McKenzie)
Date: Wed, 18 Jan 2017 16:03:34 +0000 (UTC)
Subject: [ih] Arpanet line speed
In-Reply-To: <CAHQj4CdtB6LuCU5iv85DhcM=cPK=_aDVv1pSEQXXGdLGNbU86w@mail.gmail.com>
References: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
	<CAA=duU1u37s_umFsq3pdSsLMz+=KCEUb=+wcD_BzeOkC3Zhj=Q@mail.gmail.com>
	<CAHQj4CdtB6LuCU5iv85DhcM=cPK=_aDVv1pSEQXXGdLGNbU86w@mail.gmail.com>
Message-ID: <501917129.8531321.1484755414448@mail.yahoo.com>

The initial ARPANET line from the US to the UK went via Norway and operated at 19.2kbps.? ARPA's seismic monitoring office had a 19.2kbps leased line from Washington DC to Norway for the transfer of seismic data from the NORSAR seismic array to a data processing center in DC.? This was part of the nuclear testing monitoring system used to detect Soviet underground nuclear tests.? Larry Roberts persuaded the seismic monitoring office to send the data via the ARPANET and allow the existing leased line to be incorporated into the ARPANET.? The line from NORSAR to University College London ran at the same data rate so that when no seismic data was on the transatlantic circuit the London TIP could make use of the full bandwidth.I think the British paid for the link from UCL to NORSAR, so Larry expanded ARPANET to Europe without using any money from his budget for communication links.
All of this is from memory, so the usual disclaimers about aging memories apply.
Cheers,Alex


      From: Craig Partridge <craig at tereschau.net>
 To: Andrew G. Malis <agmalis at gmail.com> 
Cc: "internet-history at postel.org" <internet-history at postel.org>; Noel Chiappa <jnc at mercury.lcs.mit.edu>
 Sent: Wednesday, January 18, 2017 10:21 AM
 Subject: Re: [ih] Arpanet line speed
   
Andy's recollection that stuff was 56 Kbps later on matches my recollection when I came on board in '83.
One thing is I don't remember what the bandwidth was on the satellite links to England.? My recollection is it washigher than 9.6Kbps but there was some oddity in integrating European and US telecom standards such that the bandwidth was different.? But this could be entirely wrong -- alas the ARPANET maps don't tell me the data rates.
Thanks!

Craig
On Wed, Jan 18, 2017 at 9:59 AM, Andrew G. Malis <agmalis at gmail.com> wrote:

By the time I came on board in '79, almost all of the links were 56 Kbps, with a few 9.6 Kbps links here and there. The 50 Kbps links had been replaced by that point.
Cheers,Andy

On Wed, Jan 18, 2017 at 9:29 AM, Noel Chiappa <jnc at mercury.lcs.mit.edu> wrote:

? ? > From: Paul Ruizendaal

? ? > - if the modem from the IMP to the Bell System was analog, the best
? ? > technology of the time was perhaps 2.4kb/s ... a speed of 50kb/s is not
? ? > a multiple of 2.4kb/s, and it would have required 21 parallel lines

If you read the 303 manual, it's clear that i) the signal between a pair of
303 modems was analog, not digital, and ii) there was a single line, with a
wide enough bandpass to carry signals of high enough frequency to carry that
bit rate - it didn't glue together a bunch of slower lines.

? ? > if the modem from the IMP to the Bell System was digital, it would most
? ? > likely have used a single channel of a T1 connection

The whole T hierarchy was just getting started then (initial deployment in the
early 1960s), and I'm not sure if it was deployed widely enough to have made
it possible to lease a T1 line from one coast to another.

Also, many of these lines would have crossed non-AT+T local phone companies
(the Bell System did not control all of the US phone system, although some
people don't realize that). The "History of the ARPANET: The First Decade"
(which I have previously pointed you at on another list), pg. III-32, says
"In the case of a circuit from UCLA to RAND ... the service would be procured
from General Telephone" - GT was the largest independent telephone company in
the US at that point. It's not clear that those local carriers would have
supported T1.


Moral of the story: when doing history, it's bad to make assumptions about
what was and wasn't possible, and about what did and did not happen. Find
contemporary documentation.

? ? ? ? Noel
_______
internet-history mailing list
internet-history at postel.org
http://mailman.postel.org/mail man/listinfo/internet-history
Contact list-owner at postel.org for assistance.



_______
internet-history mailing list
internet-history at postel.org
http://mailman.postel.org/ mailman/listinfo/internet- history
Contact list-owner at postel.org for assistance.





-- 
*****
Craig Partridge's email account for professional society activities and mailing lists.For Raytheon business, please email: craig.partridge at raytheon.com
_______
internet-history mailing list
internet-history at postel.org
http://mailman.postel.org/mailman/listinfo/internet-history
Contact list-owner at postel.org for assistance.


   
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170118/31159766/attachment.htm>

From jnc at mercury.lcs.mit.edu  Wed Jan 18 10:32:47 2017
From: jnc at mercury.lcs.mit.edu (Noel Chiappa)
Date: Wed, 18 Jan 2017 13:32:47 -0500 (EST)
Subject: [ih] Arpanet line speed
Message-ID: <20170118183247.3D04818C099@mercury.lcs.mit.edu>

    > From: "Andrew G. Malis"

    > By the time I came on board in '79, almost all of the links were 56
    > Kbps ... The 50 Kbps links had been replaced by that point.

Hmmm. I'm trying to remember when the giant rack that contained the 303
modems at MIT went away. (I should remember this, as I wound up caring for
the MIT IMPs, but I just don't recall.) Do you know any details of how the
56Kbps links worked? I.e. was there a analog modem involved, or was it
somehow provided by an underlying digital (TDMA of some sort) link, and what
kind of line was it?

Do you know why the change was made? I'm going to guess that possibly it was
cost, that whatever was used for the 56kbps was cheaper than the leased lines
used with the 303s?

	Noel


From agmalis at gmail.com  Wed Jan 18 10:56:36 2017
From: agmalis at gmail.com (Andrew G. Malis)
Date: Wed, 18 Jan 2017 13:56:36 -0500
Subject: [ih] Arpanet line speed
In-Reply-To: <20170118183247.3D04818C099@mercury.lcs.mit.edu>
References: <20170118183247.3D04818C099@mercury.lcs.mit.edu>
Message-ID: <CAA=duU2EcNNw7F0KD_0Mwgjv+x1UystR_Cx3RuVg=xXrFyf+tQ@mail.gmail.com>

Noel,

They were digital DS0s. I?m sure that they were less expensive than analog
(although I have no particular documentation of that), and most probably
more reliable as well. As I recall, with the digital circuits, the most
common cause of line outage was microwave fade.

Cheers,
Andy


On Wed, Jan 18, 2017 at 1:32 PM, Noel Chiappa <jnc at mercury.lcs.mit.edu>
wrote:

>     > From: "Andrew G. Malis"
>
>     > By the time I came on board in '79, almost all of the links were 56
>     > Kbps ... The 50 Kbps links had been replaced by that point.
>
> Hmmm. I'm trying to remember when the giant rack that contained the 303
> modems at MIT went away. (I should remember this, as I wound up caring for
> the MIT IMPs, but I just don't recall.) Do you know any details of how the
> 56Kbps links worked? I.e. was there a analog modem involved, or was it
> somehow provided by an underlying digital (TDMA of some sort) link, and
> what
> kind of line was it?
>
> Do you know why the change was made? I'm going to guess that possibly it
> was
> cost, that whatever was used for the 56kbps was cheaper than the leased
> lines
> used with the 303s?
>
>         Noel
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170118/7ea609c2/attachment.htm>

From brian.e.carpenter at gmail.com  Wed Jan 18 11:34:39 2017
From: brian.e.carpenter at gmail.com (Brian E Carpenter)
Date: Thu, 19 Jan 2017 08:34:39 +1300
Subject: [ih] Origin of "best effort"
Message-ID: <80906d2c-0359-84b3-ab61-c169efec0dd6@gmail.com>

I learnt very early on that the Internet offered a "best effort" service
for the delivery of datagrams.

Where did that meme come from, and when?

The earliest trace I found in a quick trawl was 1986 (RFC992). But RFC791
doesn't mention it, and defined TOS, such that all packets were *not* assumed
to be created equal. The 1984 Saltzer et al paper doesn't mention it either.

(RFC768 does say that UDP delivery is "not guaranteed" but that is not
the same thing as "best effort".)

The question is of interest because some analyses of network neutrality,
including a student dissertation I was reviewing yesterday, conflate the
end-to-end principle with best-effort packet delivery.

Regards
     Brian


From jack at 3kitty.org  Wed Jan 18 11:46:01 2017
From: jack at 3kitty.org (Jack Haverty)
Date: Wed, 18 Jan 2017 11:46:01 -0800
Subject: [ih] Arpanet line speed
In-Reply-To: <CAA=duU2EcNNw7F0KD_0Mwgjv+x1UystR_Cx3RuVg=xXrFyf+tQ@mail.gmail.com>
References: <20170118183247.3D04818C099@mercury.lcs.mit.edu>
	<CAA=duU2EcNNw7F0KD_0Mwgjv+x1UystR_Cx3RuVg=xXrFyf+tQ@mail.gmail.com>
Message-ID: <30ea98a6-f8fc-d221-2d37-08f0bf7ff1c5@3kitty.org>

For the historians - what those expensive "high-speed" lines actually
looked like...

I was one of the setup crew at the ICCC '72 demonstration of the ARPANET
in the ballroom of the Washington Hilton.  We had laid a false floor to
cover all the wiring between a gaggle of terminals and the TIP, put
various terminals, printers, and of course the TIP right in the middle
of the room, and were awaiting the Telco guy to come with The Line.

I expected some kind of impressive cable with esoteric connector, as
would befit the rare not-so-ordinary phone line.

When the phone guy arrived, he dragged over a ratty looking pile of thin
cable from the wall of the ballroom, and stretched it out to reach the
IMP.  It looked like what you might use to hook up your doorbell.  Or an
extension phone in your house.  Just a couple of small wires to hook up,
IIRC.  Probably the ones they had used for the last umpteen meetings in
that room where they wanted a phone.

Since it wasn't very impressive, we hid it under the false floor, duct
tape, etc. and hoped nobody would trip over it.

I remember being amazed that you could possibly get 50+ kilobits/second
over such wiring.  Of course I'm even more amazed at the gigabit/second
flowing over the tiny Ethernet wires attached to my desktop today.

/Jack

On 01/18/2017 10:56 AM, Andrew G. Malis wrote:
> Noel,
> 
> They were digital DS0s. I?m sure that they were less expensive than
> analog (although I have no particular documentation of that), and most
> probably more reliable as well. As I recall, with the digital circuits,
> the most common cause of line outage was microwave fade.
> 
> Cheers,
> Andy
> 
> 
> On Wed, Jan 18, 2017 at 1:32 PM, Noel Chiappa <jnc at mercury.lcs.mit.edu
> <mailto:jnc at mercury.lcs.mit.edu>> wrote:
> 
>         > From: "Andrew G. Malis"
> 
>         > By the time I came on board in '79, almost all of the links
>     were 56
>         > Kbps ... The 50 Kbps links had been replaced by that point.
> 
>     Hmmm. I'm trying to remember when the giant rack that contained the 303
>     modems at MIT went away. (I should remember this, as I wound up
>     caring for
>     the MIT IMPs, but I just don't recall.) Do you know any details of
>     how the
>     56Kbps links worked? I.e. was there a analog modem involved, or was it
>     somehow provided by an underlying digital (TDMA of some sort) link,
>     and what
>     kind of line was it?
> 
>     Do you know why the change was made? I'm going to guess that
>     possibly it was
>     cost, that whatever was used for the 56kbps was cheaper than the
>     leased lines
>     used with the 303s?
> 
>             Noel
> 
> 
> 
> 
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
> 


From brian.e.carpenter at gmail.com  Wed Jan 18 11:46:24 2017
From: brian.e.carpenter at gmail.com (Brian E Carpenter)
Date: Thu, 19 Jan 2017 08:46:24 +1300
Subject: [ih] Arpanet line speed
In-Reply-To: <CAHQj4CdtB6LuCU5iv85DhcM=cPK=_aDVv1pSEQXXGdLGNbU86w@mail.gmail.com>
References: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
	<CAA=duU1u37s_umFsq3pdSsLMz+=KCEUb=+wcD_BzeOkC3Zhj=Q@mail.gmail.com>
	<CAHQj4CdtB6LuCU5iv85DhcM=cPK=_aDVv1pSEQXXGdLGNbU86w@mail.gmail.com>
Message-ID: <3fd5ed61-969b-ae2b-c34c-d171e26914c4@gmail.com>

On 19/01/2017 04:21, Craig Partridge wrote:
> Andy's recollection that stuff was 56 Kbps later on matches my recollection
> when I came on board in '83.
> 
> One thing is I don't remember what the bandwidth was on the satellite links
> to England.  My recollection is it was
> higher than 9.6Kbps but there was some oddity in integrating European and
> US telecom standards such that the bandwidth was different.  But this could
> be entirely wrong -- alas the ARPANET maps don't tell me the data rates.

Google for Peter Kirstein's paper "Early Experiences with the ARPANET and INTERNET in the UK"

It was 9.6 between London and Norway and then 9.6 between Norway and the USA.

9.6 was the lowest common denominator; the US 56k standard was different from
the CCITT 64k standard used in Europe. Just as the US T1 1.5M standard was different
from the CCITT E1 2M standard some years later. This was a pain in the neck for
early transatlantic links above 9.6 (or posibly 19.2). But I suspect that in 1973
the main issue was cost.

   Brian

> 
> Thanks!
> 
> Craig
> 
> On Wed, Jan 18, 2017 at 9:59 AM, Andrew G. Malis <agmalis at gmail.com> wrote:
> 
>> By the time I came on board in '79, almost all of the links were 56 Kbps,
>> with a few 9.6 Kbps links here and there. The 50 Kbps links had been
>> replaced by that point.
>>
>> Cheers,
>> Andy
>>
>>
>> On Wed, Jan 18, 2017 at 9:29 AM, Noel Chiappa <jnc at mercury.lcs.mit.edu>
>> wrote:
>>
>>>     > From: Paul Ruizendaal
>>>
>>>     > - if the modem from the IMP to the Bell System was analog, the best
>>>     > technology of the time was perhaps 2.4kb/s ... a speed of 50kb/s is
>>> not
>>>     > a multiple of 2.4kb/s, and it would have required 21 parallel lines
>>>
>>> If you read the 303 manual, it's clear that i) the signal between a pair
>>> of
>>> 303 modems was analog, not digital, and ii) there was a single line, with
>>> a
>>> wide enough bandpass to carry signals of high enough frequency to carry
>>> that
>>> bit rate - it didn't glue together a bunch of slower lines.
>>>
>>>     > if the modem from the IMP to the Bell System was digital, it would
>>> most
>>>     > likely have used a single channel of a T1 connection
>>>
>>> The whole T hierarchy was just getting started then (initial deployment
>>> in the
>>> early 1960s), and I'm not sure if it was deployed widely enough to have
>>> made
>>> it possible to lease a T1 line from one coast to another.
>>>
>>> Also, many of these lines would have crossed non-AT+T local phone
>>> companies
>>> (the Bell System did not control all of the US phone system, although some
>>> people don't realize that). The "History of the ARPANET: The First Decade"
>>> (which I have previously pointed you at on another list), pg. III-32, says
>>> "In the case of a circuit from UCLA to RAND ... the service would be
>>> procured
>>> from General Telephone" - GT was the largest independent telephone
>>> company in
>>> the US at that point. It's not clear that those local carriers would have
>>> supported T1.
>>>
>>>
>>> Moral of the story: when doing history, it's bad to make assumptions about
>>> what was and wasn't possible, and about what did and did not happen. Find
>>> contemporary documentation.
>>>
>>>         Noel
>>> _______
>>> internet-history mailing list
>>> internet-history at postel.org
>>> http://mailman.postel.org/mailman/listinfo/internet-history
>>> Contact list-owner at postel.org for assistance.
>>>
>>
>>
>> _______
>> internet-history mailing list
>> internet-history at postel.org
>> http://mailman.postel.org/mailman/listinfo/internet-history
>> Contact list-owner at postel.org for assistance.
>>
>>
> 
> 
> 
> 
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
> 


From reed at reedmedia.net  Wed Jan 18 11:48:00 2017
From: reed at reedmedia.net (Jeremy C. Reed)
Date: Wed, 18 Jan 2017 13:48:00 -0600 (CST)
Subject: [ih] Arpanet line speed
In-Reply-To: <880BEBAA-323D-467C-821C-1A8683F4C8CC@gmail.com>
References: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
	<CAHxHggc8g4Y8n3xcoaXOtt2+nVjaqbqpReBjtpTt4=YeN25ywA@mail.gmail.com>
	<880BEBAA-323D-467C-821C-1A8683F4C8CC@gmail.com>
Message-ID: <alpine.NEB.2.20.1701181342390.17828@t1.m.reedmedia.net>

This got me thinking ... is there any list of where the original IMPs 
(and modems, etc) are located today? What museums or other places can we 
visit to see them and learn about them (in person)?

On that note, any good museums that have exhibits for the physical 
history of the internet?

Hopefully this list allows my attachment. This is an Digital IMP11a. It 
is probably around 12 years newer than the original IMPs. I took this 
photo a few years ago in a Internet Systems Consortium (ISC) warehouse 
in Redwood City, Calif.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: imp-at-isc.jpg
Type: image/jpeg
Size: 71526 bytes
Desc: 
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170118/fc9dd597/attachment.jpg>

From dave.walden.family at gmail.com  Wed Jan 18 12:51:11 2017
From: dave.walden.family at gmail.com (dave.walden.family at gmail.com)
Date: Wed, 18 Jan 2017 15:51:11 -0500
Subject: [ih] Arpanet line speed
In-Reply-To: <alpine.NEB.2.20.1701181342390.17828@t1.m.reedmedia.net>
References: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
	<CAHxHggc8g4Y8n3xcoaXOtt2+nVjaqbqpReBjtpTt4=YeN25ywA@mail.gmail.com>
	<880BEBAA-323D-467C-821C-1A8683F4C8CC@gmail.com>
	<alpine.NEB.2.20.1701181342390.17828@t1.m.reedmedia.net>
Message-ID: <14A0B29C-4B74-4675-B903-03A5575999CE@gmail.com>

 imp 1 is still at ucla and imp 10 is at the computer history museum.  i don't know about others.

Sent from my iPad

On Jan 18, 2017, at 2:48 PM, "Jeremy C. Reed" <reed at reedmedia.net> wrote:

> This got me thinking ... is there any list of where the original IMPs 
> (and modems, etc) are located today? What museums or other places can we 
> visit to see them and learn about them (in person)?
> 
> On that note, any good museums that have exhibits for the physical 
> history of the internet?
> 
> Hopefully this list allows my attachment. This is an Digital IMP11a. It 
> is probably around 12 years newer than the original IMPs. I took this 
> photo a few years ago in a Internet Systems Consortium (ISC) warehouse 
> in Redwood City, Calif.
> <imp-at-isc.jpg>
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.



From dave.walden.family at gmail.com  Wed Jan 18 12:55:18 2017
From: dave.walden.family at gmail.com (dave.walden.family at gmail.com)
Date: Wed, 18 Jan 2017 15:55:18 -0500
Subject: [ih] Arpanet line speed
In-Reply-To: <alpine.NEB.2.20.1701181342390.17828@t1.m.reedmedia.net>
References: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
	<CAHxHggc8g4Y8n3xcoaXOtt2+nVjaqbqpReBjtpTt4=YeN25ywA@mail.gmail.com>
	<880BEBAA-323D-467C-821C-1A8683F4C8CC@gmail.com>
	<alpine.NEB.2.20.1701181342390.17828@t1.m.reedmedia.net>
Message-ID: <7FDAD2AA-C62E-405E-8C9B-30F2FE652839@gmail.com>

I don't recognize the machine in the photo.

Sent from my iPad

On Jan 18, 2017, at 2:48 PM, "Jeremy C. Reed" <reed at reedmedia.net> wrote:

> This got me thinking ... is there any list of where the original IMPs 
> (and modems, etc) are located today? What museums or other places can we 
> visit to see them and learn about them (in person)?
> 
> On that note, any good museums that have exhibits for the physical 
> history of the internet?
> 
> Hopefully this list allows my attachment. This is an Digital IMP11a. It 
> is probably around 12 years newer than the original IMPs. I took this 
> photo a few years ago in a Internet Systems Consortium (ISC) warehouse 
> in Redwood City, Calif.
> <imp-at-isc.jpg>
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.


From amckenzie3 at yahoo.com  Wed Jan 18 12:55:48 2017
From: amckenzie3 at yahoo.com (Alex McKenzie)
Date: Wed, 18 Jan 2017 20:55:48 +0000 (UTC)
Subject: [ih] Arpanet line speed
In-Reply-To: <alpine.NEB.2.20.1701181342390.17828@t1.m.reedmedia.net>
References: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
	<CAHxHggc8g4Y8n3xcoaXOtt2+nVjaqbqpReBjtpTt4=YeN25ywA@mail.gmail.com>
	<880BEBAA-323D-467C-821C-1A8683F4C8CC@gmail.com>
	<alpine.NEB.2.20.1701181342390.17828@t1.m.reedmedia.net>
Message-ID: <1770679276.8897154.1484772948401@mail.yahoo.com>

Jeremy,
The IMP 11a is NOT an Interface Message Processor, and was not an ARPANET? (or ARPANET clone) switch.? It MAY have been a part of ARPA's packet radio experiments, but I am not confident that it had anything to do with packet switching.
Cheers,Alex McKenzie


      From: Jeremy C. Reed <reed at reedmedia.net>
 To: internet history <internet-history at postel.org> 
 Sent: Wednesday, January 18, 2017 2:48 PM
 Subject: Re: [ih] Arpanet line speed
   
This got me thinking ... is there any list of where the original IMPs 
(and modems, etc) are located today? What museums or other places can we 
visit to see them and learn about them (in person)?

On that note, any good museums that have exhibits for the physical 
history of the internet?

Hopefully this list allows my attachment. This is an Digital IMP11a. It 
is probably around 12 years newer than the original IMPs. I took this 
photo a few years ago in a Internet Systems Consortium (ISC) warehouse 
in Redwood City, Calif.
_______
internet-history mailing list
internet-history at postel.org
http://mailman.postel.org/mailman/listinfo/internet-history
Contact list-owner at postel.org for assistance.


   
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170118/5fa61094/attachment.htm>

From vint at google.com  Wed Jan 18 12:55:58 2017
From: vint at google.com (Vint Cerf)
Date: Wed, 18 Jan 2017 15:55:58 -0500
Subject: [ih] Arpanet line speed
In-Reply-To: <3fd5ed61-969b-ae2b-c34c-d171e26914c4@gmail.com>
References: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
	<CAA=duU1u37s_umFsq3pdSsLMz+=KCEUb=+wcD_BzeOkC3Zhj=Q@mail.gmail.com>
	<CAHQj4CdtB6LuCU5iv85DhcM=cPK=_aDVv1pSEQXXGdLGNbU86w@mail.gmail.com>
	<3fd5ed61-969b-ae2b-c34c-d171e26914c4@gmail.com>
Message-ID: <CAHxHggc9bZeOu=LfqqLsJzJBn7bVO9nvy9RrMBkknLPyyM1bpg@mail.gmail.com>

1. the internal satellite link that connected the US to Norway was a shared
9.6 kib/s link. ARPANET and the NORSAR/SEISMO systems shared that channel.
I don't think that ever got higher until it was retired and replaced in
1982 when Europe was forced to go over TCP/IP on the shared SATNET 64 kb/s
link (later expanded to 128kb/s). The rest of the ARPANET hosts ran a mix
of NCP and TCP until January 1, 1983 when almost all hosts switched to
TCP/IP. There were a couple of exceptions that were given a few months'
reprieve.

2. The ARPANET leased lines were 50 kb/s until they were replaced by 64
kb/s DS0's except that 1 bit of 8 was taken as a framing bit, which stole 8
kb/s from 64 leaving 56 kb/s for data carrying in the US.

at least that is what I seem to remember.

v


On Wed, Jan 18, 2017 at 2:46 PM, Brian E Carpenter <
brian.e.carpenter at gmail.com> wrote:

> On 19/01/2017 04:21, Craig Partridge wrote:
> > Andy's recollection that stuff was 56 Kbps later on matches my
> recollection
> > when I came on board in '83.
> >
> > One thing is I don't remember what the bandwidth was on the satellite
> links
> > to England.  My recollection is it was
> > higher than 9.6Kbps but there was some oddity in integrating European and
> > US telecom standards such that the bandwidth was different.  But this
> could
> > be entirely wrong -- alas the ARPANET maps don't tell me the data rates.
>
> Google for Peter Kirstein's paper "Early Experiences with the ARPANET and
> INTERNET in the UK"
>
> It was 9.6 between London and Norway and then 9.6 between Norway and the
> USA.
>
> 9.6 was the lowest common denominator; the US 56k standard was different
> from
> the CCITT 64k standard used in Europe. Just as the US T1 1.5M standard was
> different
> from the CCITT E1 2M standard some years later. This was a pain in the
> neck for
> early transatlantic links above 9.6 (or posibly 19.2). But I suspect that
> in 1973
> the main issue was cost.
>
>    Brian
>
> >
> > Thanks!
> >
> > Craig
> >
> > On Wed, Jan 18, 2017 at 9:59 AM, Andrew G. Malis <agmalis at gmail.com>
> wrote:
> >
> >> By the time I came on board in '79, almost all of the links were 56
> Kbps,
> >> with a few 9.6 Kbps links here and there. The 50 Kbps links had been
> >> replaced by that point.
> >>
> >> Cheers,
> >> Andy
> >>
> >>
> >> On Wed, Jan 18, 2017 at 9:29 AM, Noel Chiappa <jnc at mercury.lcs.mit.edu>
> >> wrote:
> >>
> >>>     > From: Paul Ruizendaal
> >>>
> >>>     > - if the modem from the IMP to the Bell System was analog, the
> best
> >>>     > technology of the time was perhaps 2.4kb/s ... a speed of 50kb/s
> is
> >>> not
> >>>     > a multiple of 2.4kb/s, and it would have required 21 parallel
> lines
> >>>
> >>> If you read the 303 manual, it's clear that i) the signal between a
> pair
> >>> of
> >>> 303 modems was analog, not digital, and ii) there was a single line,
> with
> >>> a
> >>> wide enough bandpass to carry signals of high enough frequency to carry
> >>> that
> >>> bit rate - it didn't glue together a bunch of slower lines.
> >>>
> >>>     > if the modem from the IMP to the Bell System was digital, it
> would
> >>> most
> >>>     > likely have used a single channel of a T1 connection
> >>>
> >>> The whole T hierarchy was just getting started then (initial deployment
> >>> in the
> >>> early 1960s), and I'm not sure if it was deployed widely enough to have
> >>> made
> >>> it possible to lease a T1 line from one coast to another.
> >>>
> >>> Also, many of these lines would have crossed non-AT+T local phone
> >>> companies
> >>> (the Bell System did not control all of the US phone system, although
> some
> >>> people don't realize that). The "History of the ARPANET: The First
> Decade"
> >>> (which I have previously pointed you at on another list), pg. III-32,
> says
> >>> "In the case of a circuit from UCLA to RAND ... the service would be
> >>> procured
> >>> from General Telephone" - GT was the largest independent telephone
> >>> company in
> >>> the US at that point. It's not clear that those local carriers would
> have
> >>> supported T1.
> >>>
> >>>
> >>> Moral of the story: when doing history, it's bad to make assumptions
> about
> >>> what was and wasn't possible, and about what did and did not happen.
> Find
> >>> contemporary documentation.
> >>>
> >>>         Noel
> >>> _______
> >>> internet-history mailing list
> >>> internet-history at postel.org
> >>> http://mailman.postel.org/mailman/listinfo/internet-history
> >>> Contact list-owner at postel.org for assistance.
> >>>
> >>
> >>
> >> _______
> >> internet-history mailing list
> >> internet-history at postel.org
> >> http://mailman.postel.org/mailman/listinfo/internet-history
> >> Contact list-owner at postel.org for assistance.
> >>
> >>
> >
> >
> >
> >
> > _______
> > internet-history mailing list
> > internet-history at postel.org
> > http://mailman.postel.org/mailman/listinfo/internet-history
> > Contact list-owner at postel.org for assistance.
> >
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
>



-- 
New postal address:
Google
1875 Explorer Street, 10th Floor
Reston, VA 20190
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170118/34361744/attachment.htm>

From bob.hinden at gmail.com  Wed Jan 18 13:10:37 2017
From: bob.hinden at gmail.com (Bob Hinden)
Date: Wed, 18 Jan 2017 13:10:37 -0800
Subject: [ih] Arpanet line speed
In-Reply-To: <alpine.NEB.2.20.1701181342390.17828@t1.m.reedmedia.net>
References: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
	<CAHxHggc8g4Y8n3xcoaXOtt2+nVjaqbqpReBjtpTt4=YeN25ywA@mail.gmail.com>
	<880BEBAA-323D-467C-821C-1A8683F4C8CC@gmail.com>
	<alpine.NEB.2.20.1701181342390.17828@t1.m.reedmedia.net>
Message-ID: <A66A053A-3456-4A98-89AC-BEC1CCA749DB@gmail.com>

Jeremy,

The device in the picture is not an IMP.  I think it?s a DEC designed interface to an IMP.  Guessing for a PDP-10.  I had a discussion with Paul Vixe when he as at ISC about this.  The thought it was an IMP :-)

There is an IMP in the Computer History Museum in Mt. View.  It?s part of their exhibits that include Arpanet and Internet.  Worth seeing!  See http://www.computerhistory.org 

Bob


> On Jan 18, 2017, at 11:48 AM, Jeremy C. Reed <reed at reedmedia.net> wrote:
> 
> This got me thinking ... is there any list of where the original IMPs 
> (and modems, etc) are located today? What museums or other places can we 
> visit to see them and learn about them (in person)?
> 
> On that note, any good museums that have exhibits for the physical 
> history of the internet?
> 
> Hopefully this list allows my attachment. This is an Digital IMP11a. It 
> is probably around 12 years newer than the original IMPs. I took this 
> photo a few years ago in a Internet Systems Consortium (ISC) warehouse 
> in Redwood City, Calif.<imp-at-isc.jpg>_______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.




From vint at google.com  Wed Jan 18 13:52:24 2017
From: vint at google.com (Vint Cerf)
Date: Wed, 18 Jan 2017 16:52:24 -0500
Subject: [ih] Arpanet line speed
In-Reply-To: <A66A053A-3456-4A98-89AC-BEC1CCA749DB@gmail.com>
References: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
	<CAHxHggc8g4Y8n3xcoaXOtt2+nVjaqbqpReBjtpTt4=YeN25ywA@mail.gmail.com>
	<880BEBAA-323D-467C-821C-1A8683F4C8CC@gmail.com>
	<alpine.NEB.2.20.1701181342390.17828@t1.m.reedmedia.net>
	<A66A053A-3456-4A98-89AC-BEC1CCA749DB@gmail.com>
Message-ID: <CAHxHggfhwMhg63c1LVM4e0M03Gr6wE0z2YtJvK4PHSVfQ84TPg@mail.gmail.com>

the Smithsonian may also have an IMP

v


On Wed, Jan 18, 2017 at 4:10 PM, Bob Hinden <bob.hinden at gmail.com> wrote:

> Jeremy,
>
> The device in the picture is not an IMP.  I think it?s a DEC designed
> interface to an IMP.  Guessing for a PDP-10.  I had a discussion with Paul
> Vixe when he as at ISC about this.  The thought it was an IMP :-)
>
> There is an IMP in the Computer History Museum in Mt. View.  It?s part of
> their exhibits that include Arpanet and Internet.  Worth seeing!  See
> http://www.computerhistory.org
>
> Bob
>
>
> > On Jan 18, 2017, at 11:48 AM, Jeremy C. Reed <reed at reedmedia.net> wrote:
> >
> > This got me thinking ... is there any list of where the original IMPs
> > (and modems, etc) are located today? What museums or other places can we
> > visit to see them and learn about them (in person)?
> >
> > On that note, any good museums that have exhibits for the physical
> > history of the internet?
> >
> > Hopefully this list allows my attachment. This is an Digital IMP11a. It
> > is probably around 12 years newer than the original IMPs. I took this
> > photo a few years ago in a Internet Systems Consortium (ISC) warehouse
> > in Redwood City, Calif.<imp-at-isc.jpg>_______
> > internet-history mailing list
> > internet-history at postel.org
> > http://mailman.postel.org/mailman/listinfo/internet-history
> > Contact list-owner at postel.org for assistance.
>
>
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
>



-- 
New postal address:
Google
1875 Explorer Street, 10th Floor
Reston, VA 20190
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170118/099acecb/attachment.htm>

From jack at 3kitty.org  Wed Jan 18 14:16:03 2017
From: jack at 3kitty.org (Jack Haverty)
Date: Wed, 18 Jan 2017 14:16:03 -0800
Subject: [ih] Arpanet line speed
In-Reply-To: <alpine.NEB.2.20.1701181342390.17828@t1.m.reedmedia.net>
References: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
	<CAHxHggc8g4Y8n3xcoaXOtt2+nVjaqbqpReBjtpTt4=YeN25ywA@mail.gmail.com>
	<880BEBAA-323D-467C-821C-1A8683F4C8CC@gmail.com>
	<alpine.NEB.2.20.1701181342390.17828@t1.m.reedmedia.net>
Message-ID: <e4f581ef-3df9-463f-97d2-3c7a6ea85191@3kitty.org>

That's an IMP interface, which was a peripheral device for the PDP-11
computers.  One of those was on the PDP-11/40 in a lab at BBN in Fall
1977, so it had become a commercial product before then.  It connected
between an IMP port, following the specifications in BBN 1822, and the
PDP-11 on the Unibus IIRC.  It didn't do much processing - just
transferred the ARPANET "messages" to and from the IMP and the computer
memory.

I wrote the first TCP for Unix, and that's how those IP packets got in
and out of the machine.  I spent many hours trying to figure out what
was going on by looking at those lights...

/Jack


On 01/18/2017 11:48 AM, Jeremy C. Reed wrote:
> This got me thinking ... is there any list of where the original IMPs 
> (and modems, etc) are located today? What museums or other places can we 
> visit to see them and learn about them (in person)?
> 
> On that note, any good museums that have exhibits for the physical 
> history of the internet?
> 
> Hopefully this list allows my attachment. This is an Digital IMP11a. It 
> is probably around 12 years newer than the original IMPs. I took this 
> photo a few years ago in a Internet Systems Consortium (ISC) warehouse 
> in Redwood City, Calif.
> 
> 
> 
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
> 


From casner at acm.org  Wed Jan 18 16:02:04 2017
From: casner at acm.org (Stephen Casner)
Date: Wed, 18 Jan 2017 16:02:04 -0800 (PST)
Subject: [ih] PDP11 ARPANET interfaces (was: Arpanet line speed)
In-Reply-To: <e4f581ef-3df9-463f-97d2-3c7a6ea85191@3kitty.org>
References: <20170118142923.A959B18C09A@mercury.lcs.mit.edu>
	<CAHxHggc8g4Y8n3xcoaXOtt2+nVjaqbqpReBjtpTt4=YeN25ywA@mail.gmail.com>
	<880BEBAA-323D-467C-821C-1A8683F4C8CC@gmail.com>
	<alpine.NEB.2.20.1701181342390.17828@t1.m.reedmedia.net>
	<e4f581ef-3df9-463f-97d2-3c7a6ea85191@3kitty.org>
Message-ID: <alpine.OSX.2.20.1701181545070.98180@auge.attlocal.net>

I corroborate Jack's description.  The IMP11a was DEC's implementation
of an interface for a PDP11 to connect to the ARPANET IMP.  Two
predecessor PDP11 interfaces were the ANTS interface from the
University of Illinois and a variant of that design made at USC
Information Sciences Institute.  My first task when I joined ISI in
1973 was to write diagnostic software for it.  All three interfaces
were provided DMA transfer into and out of memory via the UNIBUS.  The
only function you might call processing was an option to swap bytes
going into or out of the 16-bit words of PDP11 memory (since the PDP11
was a little-endian inhabiting a landscape which at that time
consisted of mostly big-endians)

The ISI PDP11 ARPANET Interface is featured in action in the opening
scene of the digital voice converencing movie we made in 1978.

https://www.youtube.com/watch?v=MGat1jRQ_SM

                                                        -- Steve

On Wed, 18 Jan 2017, Jack Haverty wrote:

> That's an IMP interface, which was a peripheral device for the PDP-11
> computers.  One of those was on the PDP-11/40 in a lab at BBN in Fall
> 1977, so it had become a commercial product before then.  It connected
> between an IMP port, following the specifications in BBN 1822, and the
> PDP-11 on the Unibus IIRC.  It didn't do much processing - just
> transferred the ARPANET "messages" to and from the IMP and the computer
> memory.
>
> I wrote the first TCP for Unix, and that's how those IP packets got in
> and out of the machine.  I spent many hours trying to figure out what
> was going on by looking at those lights...
>
> /Jack
>
>
> On 01/18/2017 11:48 AM, Jeremy C. Reed wrote:
> > This got me thinking ... is there any list of where the original IMPs
> > (and modems, etc) are located today? What museums or other places can we
> > visit to see them and learn about them (in person)?
> >
> > On that note, any good museums that have exhibits for the physical
> > history of the internet?
> >
> > Hopefully this list allows my attachment. This is an Digital IMP11a. It
> > is probably around 12 years newer than the original IMPs. I took this
> > photo a few years ago in a Internet Systems Consortium (ISC) warehouse
> > in Redwood City, Calif.


From jeanjour at comcast.net  Wed Jan 18 17:35:38 2017
From: jeanjour at comcast.net (John Day)
Date: Wed, 18 Jan 2017 20:35:38 -0500
Subject: [ih] Origin of "best effort"
In-Reply-To: <80906d2c-0359-84b3-ab61-c169efec0dd6@gmail.com>
References: <80906d2c-0359-84b3-ab61-c169efec0dd6@gmail.com>
Message-ID: <255A069C-D3A9-4435-A835-2216EFBBD815@comcast.net>

I first heard it from Pouzin or Zimmermann. CYCLADES was the original ?best effort? network.  Pouzin?s argument for a datagram network was that the hosts would check to ensure that all of the data was received, so the network didn?t have to be perfect. It only had to make a "best effort? so that it wasn?t too expensive to recover the errors.

Take care,
John

> On Jan 18, 2017, at 14:34, Brian E Carpenter <brian.e.carpenter at gmail.com> wrote:
> 
> I learnt very early on that the Internet offered a "best effort" service
> for the delivery of datagrams.
> 
> Where did that meme come from, and when?
> 
> The earliest trace I found in a quick trawl was 1986 (RFC992). But RFC791
> doesn't mention it, and defined TOS, such that all packets were *not* assumed
> to be created equal. The 1984 Saltzer et al paper doesn't mention it either.
> 
> (RFC768 does say that UDP delivery is "not guaranteed" but that is not
> the same thing as "best effort".)
> 
> The question is of interest because some analyses of network neutrality,
> including a student dissertation I was reviewing yesterday, conflate the
> end-to-end principle with best-effort packet delivery.
> 
> Regards
>     Brian
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.




From galmes at tamu.edu  Thu Jan 19 11:09:43 2017
From: galmes at tamu.edu (Guy Almes)
Date: Thu, 19 Jan 2017 14:09:43 -0500
Subject: [ih] Origin of "best effort"
In-Reply-To: <80906d2c-0359-84b3-ab61-c169efec0dd6@gmail.com>
References: <80906d2c-0359-84b3-ab61-c169efec0dd6@gmail.com>
Message-ID: <162fc808-106b-7bcb-cbe5-215dd166296f@tamu.edu>

Hi Brian,
   This is actually an interesting question.
   During the 1990s we had a discussion in which this phrase came up, 
used in the now-typical networking sense.
   One colleague in the conversation, a lawyer, who had done work in the 
construction industry, found the usage very odd because, he said, in 
contracts within that industry the phrase had a specific meaning and it 
obliged a person/company to a very very high standard of "best effort". 
In that context, for example, it might oblige a company to spend 
money/effort to a degree that would keep a promise but ruin any hope for 
making money in the deal.  He noted that we computer engineers were 
using the term in an almost opposite (and, to him, an ironic) sense, 
viz., "do whatever is normal, but *not* heroic, and if it works, good 
and if it doesn't, don't worry about it".

   If the phrase is connected that older "contract language" usage, it 
would indeed be interesting to see how it came to have almost a reversal 
or sense.

	-- Guy

On 1/18/17 2:34 PM, Brian E Carpenter wrote:
> I learnt very early on that the Internet offered a "best effort" service
> for the delivery of datagrams.
>
> Where did that meme come from, and when?
>
> The earliest trace I found in a quick trawl was 1986 (RFC992). But RFC791
> doesn't mention it, and defined TOS, such that all packets were *not* assumed
> to be created equal. The 1984 Saltzer et al paper doesn't mention it either.
>
> (RFC768 does say that UDP delivery is "not guaranteed" but that is not
> the same thing as "best effort".)
>
> The question is of interest because some analyses of network neutrality,
> including a student dissertation I was reviewing yesterday, conflate the
> end-to-end principle with best-effort packet delivery.
>
> Regards
>      Brian
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
>


From jeanjour at comcast.net  Thu Jan 19 11:35:01 2017
From: jeanjour at comcast.net (John Day)
Date: Thu, 19 Jan 2017 14:35:01 -0500
Subject: [ih] Origin of "best effort"
In-Reply-To: <162fc808-106b-7bcb-cbe5-215dd166296f@tamu.edu>
References: <80906d2c-0359-84b3-ab61-c169efec0dd6@gmail.com>
	<162fc808-106b-7bcb-cbe5-215dd166296f@tamu.edu>
Message-ID: <5AED9FBC-9738-411D-B6C0-5DF7D6DD7972@comcast.net>

For most of my life, the term ?to make a best effort? had more the latter meaning that the former. It was just a colloquial phrase. I wouldn?t call it slang but similar.  As I said, in networking the phrase originates with Pouzin and CYCLADES and the invention of the datagram. That would put it in the early 1970s.

John

> On Jan 19, 2017, at 14:09, Guy Almes <galmes at tamu.edu> wrote:
> 
> Hi Brian,
>   This is actually an interesting question.
>   During the 1990s we had a discussion in which this phrase came up, 
> used in the now-typical networking sense.
>   One colleague in the conversation, a lawyer, who had done work in the 
> construction industry, found the usage very odd because, he said, in 
> contracts within that industry the phrase had a specific meaning and it 
> obliged a person/company to a very very high standard of "best effort". 
> In that context, for example, it might oblige a company to spend 
> money/effort to a degree that would keep a promise but ruin any hope for 
> making money in the deal.  He noted that we computer engineers were 
> using the term in an almost opposite (and, to him, an ironic) sense, 
> viz., "do whatever is normal, but *not* heroic, and if it works, good 
> and if it doesn't, don't worry about it".
> 
>   If the phrase is connected that older "contract language" usage, it 
> would indeed be interesting to see how it came to have almost a reversal 
> or sense.
> 
> 	-- Guy
> 
> On 1/18/17 2:34 PM, Brian E Carpenter wrote:
>> I learnt very early on that the Internet offered a "best effort" service
>> for the delivery of datagrams.
>> 
>> Where did that meme come from, and when?
>> 
>> The earliest trace I found in a quick trawl was 1986 (RFC992). But RFC791
>> doesn't mention it, and defined TOS, such that all packets were *not* assumed
>> to be created equal. The 1984 Saltzer et al paper doesn't mention it either.
>> 
>> (RFC768 does say that UDP delivery is "not guaranteed" but that is not
>> the same thing as "best effort".)
>> 
>> The question is of interest because some analyses of network neutrality,
>> including a student dissertation I was reviewing yesterday, conflate the
>> end-to-end principle with best-effort packet delivery.
>> 
>> Regards
>>     Brian
>> _______
>> internet-history mailing list
>> internet-history at postel.org
>> http://mailman.postel.org/mailman/listinfo/internet-history
>> Contact list-owner at postel.org for assistance.
>> 
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.




From brian.e.carpenter at gmail.com  Thu Jan 19 17:07:57 2017
From: brian.e.carpenter at gmail.com (Brian E Carpenter)
Date: Fri, 20 Jan 2017 14:07:57 +1300
Subject: [ih] Origin of "best effort"
In-Reply-To: <5AED9FBC-9738-411D-B6C0-5DF7D6DD7972@comcast.net>
References: <80906d2c-0359-84b3-ab61-c169efec0dd6@gmail.com>
	<162fc808-106b-7bcb-cbe5-215dd166296f@tamu.edu>
	<5AED9FBC-9738-411D-B6C0-5DF7D6DD7972@comcast.net>
Message-ID: <bf4b376d-c98c-5ca1-91b8-a44b44c24340@gmail.com>

http://mailman.postel.org/pipermail/internet-history/2010-June/001360.html
offered another take. However, I haven't actually found a contemporary
publication by Pouzin that uses the phrase. Secondary sources cite his
1974 paper "CIGALE, the packet switching machine of the CYCLADES". That
has a nice summary of the e2e argument 10 years before Saltzer et al:
"However node and line failures coupled with adaptive routing may result
in packets being lost or duplicated. Consequently, some control mechanism
is necessary to catch this type of error. It can only be done as part of
a transmission procedure between a pair of correspondents."
It also covers TOS: "...it is clear that various traffic classes might be
accommodated, e.g. a shortest delay class,..."
It's a brilliant paper for its date, but it does not describe "best effort"
service.

In fact it seems to me that most work at that time (ALOHA, for example)
worried extensively about errors and repeating faulty packets at layer 2.
Even CSMA/CD and WiFi do that to this day. So even admiring Pouzin's
contribution, I'm still puzzled about when "best effort" (which is not an
empty phrase) entered the formal discourse. It doesn't appear in any of
Pouzin's publications from that era in the IEEE and ACM libraries, and
I can't find it in the 1979 edition of Davies et al "Computer Networks
and Their Protocols", which cites Pouzin's work.

Regards
   Brian

On 20/01/2017 08:35, John Day wrote:
> For most of my life, the term ?to make a best effort? had more the latter meaning that the former. It was just a colloquial phrase. I wouldn?t call it slang but similar.  As I said, in networking the phrase originates with Pouzin and CYCLADES and the invention of the datagram. That would put it in the early 1970s.
> 
> John
> 
>> On Jan 19, 2017, at 14:09, Guy Almes <galmes at tamu.edu> wrote:
>>
>> Hi Brian,
>>   This is actually an interesting question.
>>   During the 1990s we had a discussion in which this phrase came up, 
>> used in the now-typical networking sense.
>>   One colleague in the conversation, a lawyer, who had done work in the 
>> construction industry, found the usage very odd because, he said, in 
>> contracts within that industry the phrase had a specific meaning and it 
>> obliged a person/company to a very very high standard of "best effort". 
>> In that context, for example, it might oblige a company to spend 
>> money/effort to a degree that would keep a promise but ruin any hope for 
>> making money in the deal.  He noted that we computer engineers were 
>> using the term in an almost opposite (and, to him, an ironic) sense, 
>> viz., "do whatever is normal, but *not* heroic, and if it works, good 
>> and if it doesn't, don't worry about it".
>>
>>   If the phrase is connected that older "contract language" usage, it 
>> would indeed be interesting to see how it came to have almost a reversal 
>> or sense.
>>
>> 	-- Guy
>>
>> On 1/18/17 2:34 PM, Brian E Carpenter wrote:
>>> I learnt very early on that the Internet offered a "best effort" service
>>> for the delivery of datagrams.
>>>
>>> Where did that meme come from, and when?
>>>
>>> The earliest trace I found in a quick trawl was 1986 (RFC992). But RFC791
>>> doesn't mention it, and defined TOS, such that all packets were *not* assumed
>>> to be created equal. The 1984 Saltzer et al paper doesn't mention it either.
>>>
>>> (RFC768 does say that UDP delivery is "not guaranteed" but that is not
>>> the same thing as "best effort".)
>>>
>>> The question is of interest because some analyses of network neutrality,
>>> including a student dissertation I was reviewing yesterday, conflate the
>>> end-to-end principle with best-effort packet delivery.
>>>
>>> Regards
>>>     Brian
>>> _______
>>> internet-history mailing list
>>> internet-history at postel.org
>>> http://mailman.postel.org/mailman/listinfo/internet-history
>>> Contact list-owner at postel.org for assistance.
>>>
>> _______
>> internet-history mailing list
>> internet-history at postel.org
>> http://mailman.postel.org/mailman/listinfo/internet-history
>> Contact list-owner at postel.org for assistance.
> 
> 




From joly at punkcast.com  Thu Jan 19 20:58:22 2017
From: joly at punkcast.com (Joly MacFie)
Date: Thu, 19 Jan 2017 23:58:22 -0500
Subject: [ih] VIDEO - History of Computer Networks in Canada
Message-ID: <CAM9VJk1jE_fdNgXEaaT4JGmTe1Qf=6FR2Nt-6wKbyxE6ede+SA@mail.gmail.com>

At the State of our Networks 2017 Event in Toronto on Jan 14, 2017. Zbigniew
Stachniak, Associate Professor in the Department of Electrical Engineering
and Computer Science at York University, Toronto, gave a brief presentation
 - History of Computer Networks in Canada.

My colleague Glenn McKnight was there with his video camera, and has posted
his recording at https://www.youtube.com/watch?v=f9utcVcAVCQ

Mentioned are plans for a swank-looking Computer Museum at YorkU.
http://www.cs.yorku.ca/museum/

j

-- 
---------------------------------------------------------------
Joly MacFie  218 565 9365 Skype:punkcast
--------------------------------------------------------------
-
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170119/b51662e7/attachment.htm>

From pnr at planet.nl  Fri Jan 20 07:01:53 2017
From: pnr at planet.nl (Paul Ruizendaal)
Date: Fri, 20 Jan 2017 16:01:53 +0100
Subject: [ih] Autodin II
Message-ID: <8F90431E-E2C2-4E99-A684-4A348C2E1A58@planet.nl>


I'm currently retyping the Wingfield 1978/79 TCP/IP implementation (written for Unix V6/V7, in C); the relevant spec appears to be IEN54 & 55

This TCP/IP stack has a significant portion of code dealing with security, precedence and user groups (TCC's). From the code it seems that this is related to facilities offered by the AUTODIN II network. IEN55 mentions in its history section that some work on TCP2 was done for AUTODIN II in 1976.

A quick glance at Google results suggests that AUTODIN II was a failing project, cancelled in 1982. However, it must have been seen as viable and relevant in the '76-'79 timeframe.

What is the connection between TCP/IP work in the 70's and Autodin II? If it has been discussed before, I would much appreciate a link.

Thanks,

Paul






From mfidelman at meetinghouse.net  Fri Jan 20 09:35:42 2017
From: mfidelman at meetinghouse.net (Miles Fidelman)
Date: Fri, 20 Jan 2017 12:35:42 -0500
Subject: [ih] Autodin II
In-Reply-To: <8F90431E-E2C2-4E99-A684-4A348C2E1A58@planet.nl>
References: <8F90431E-E2C2-4E99-A684-4A348C2E1A58@planet.nl>
Message-ID: <7e5ee186-265d-a9aa-c1d0-e790e954b4d1@meetinghouse.net>

AUTODIN II was a planned replacement for AUTODIN - the DoD message 
switching system.

As a result of some well-played techno-politics, a "bakeoff" was held 
between ARPANET technology and the technology proposed for AUTODIN II, 
leading to award of the contracts for the Defense Data Network (the 
MILNET, split off from the ARPANET, and several separate classified 
networks).  AUTODIN II was cancelled.

There are others on the list who were part of this and can probably 
supply some details.

Miles Fidelman


On 1/20/17 10:01 AM, Paul Ruizendaal wrote:
> I'm currently retyping the Wingfield 1978/79 TCP/IP implementation (written for Unix V6/V7, in C); the relevant spec appears to be IEN54 & 55
>
> This TCP/IP stack has a significant portion of code dealing with security, precedence and user groups (TCC's). From the code it seems that this is related to facilities offered by the AUTODIN II network. IEN55 mentions in its history section that some work on TCP2 was done for AUTODIN II in 1976.
>
> A quick glance at Google results suggests that AUTODIN II was a failing project, cancelled in 1982. However, it must have been seen as viable and relevant in the '76-'79 timeframe.
>
> What is the connection between TCP/IP work in the 70's and Autodin II? If it has been discussed before, I would much appreciate a link.
>
> Thanks,
>
> Paul
>
>
>
>
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.

-- 
In theory, there is no difference between theory and practice.
In practice, there is.  .... Yogi Berra



From mbgreen at seas.upenn.edu  Fri Jan 20 09:50:19 2017
From: mbgreen at seas.upenn.edu (Michael Greenwald)
Date: Fri, 20 Jan 2017 09:50:19 -0800
Subject: [ih] Origin of "best effort"
In-Reply-To: <162fc808-106b-7bcb-cbe5-215dd166296f@tamu.edu>
References: <80906d2c-0359-84b3-ab61-c169efec0dd6@gmail.com>
	<162fc808-106b-7bcb-cbe5-215dd166296f@tamu.edu>
Message-ID: <afb3ed5b2cda43766d50549d1b8ad852@seas.upenn.edu>

Are we asking about the specific phrase "best effort"?
Or the notion that IP doesn't guarantee preserving
packet sequence, or even guarantee delivery?

I assume not the latter, because the Catenet paper said:

"It is very important to note that it is explicitly assumed that
datagrams are not necessarily kept in the same sequence on
exiting a network as when they entered.  Furthermore, it is
assumed that datagrams may be lost or even duplicated within the
network.  It is left up to higher level protocols in the catenet
model to recover from any problems these assumptions may
introduce."

And RFC791 restates this as:

"The internet protocol does not provide a reliable communication
facility.  There are no acknowledgments either end-to-end or
hop-by-hop.  There is no error control for data, only a header
checksum.  There are no retransmissions.  There is no flow control."

To my mind this explains what was later more loosely referred to
as "best effort".  Is there any reason to believe it was used as
a term of art, rather than a shorthand description of the 2 paragraphs
I quoted? (I know that when I first heard "best effort" I
interpreted it as a summary, but that just describes my reaction,
not the intent of whoever used it).

On 2017-01-19 11:09, Guy Almes wrote:
> Hi Brian,
>    This is actually an interesting question.
>    During the 1990s we had a discussion in which this phrase came up,
> used in the now-typical networking sense.
>    One colleague in the conversation, a lawyer, who had done work in 
> the
> construction industry, found the usage very odd because, he said, in
> contracts within that industry the phrase had a specific meaning and it
> obliged a person/company to a very very high standard of "best effort".
> In that context, for example, it might oblige a company to spend
> money/effort to a degree that would keep a promise but ruin any hope 
> for
> making money in the deal.  He noted that we computer engineers were
> using the term in an almost opposite (and, to him, an ironic) sense,
> viz., "do whatever is normal, but *not* heroic, and if it works, good
> and if it doesn't, don't worry about it".
> 
>    If the phrase is connected that older "contract language" usage, it
> would indeed be interesting to see how it came to have almost a 
> reversal
> or sense.
> 
> 	-- Guy
> 
> On 1/18/17 2:34 PM, Brian E Carpenter wrote:
>> I learnt very early on that the Internet offered a "best effort" 
>> service
>> for the delivery of datagrams.
>> 
>> Where did that meme come from, and when?
>> 
>> The earliest trace I found in a quick trawl was 1986 (RFC992). But 
>> RFC791
>> doesn't mention it, and defined TOS, such that all packets were *not* 
>> assumed
>> to be created equal. The 1984 Saltzer et al paper doesn't mention it 
>> either.
>> 
>> (RFC768 does say that UDP delivery is "not guaranteed" but that is not
>> the same thing as "best effort".)
>> 
>> The question is of interest because some analyses of network 
>> neutrality,
>> including a student dissertation I was reviewing yesterday, conflate 
>> the
>> end-to-end principle with best-effort packet delivery.
>> 
>> Regards
>>      Brian
>> _______
>> internet-history mailing list
>> internet-history at postel.org
>> http://mailman.postel.org/mailman/listinfo/internet-history
>> Contact list-owner at postel.org for assistance.
>> 
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.


From brian.e.carpenter at gmail.com  Fri Jan 20 11:28:16 2017
From: brian.e.carpenter at gmail.com (Brian E Carpenter)
Date: Sat, 21 Jan 2017 08:28:16 +1300
Subject: [ih] Origin of "best effort"
In-Reply-To: <afb3ed5b2cda43766d50549d1b8ad852@seas.upenn.edu>
References: <80906d2c-0359-84b3-ab61-c169efec0dd6@gmail.com>
	<162fc808-106b-7bcb-cbe5-215dd166296f@tamu.edu>
	<afb3ed5b2cda43766d50549d1b8ad852@seas.upenn.edu>
Message-ID: <3621d309-2d4f-1f3a-a8cc-bf9435a8cd74@gmail.com>

On 21/01/2017 06:50, Michael Greenwald wrote:
> Are we asking about the specific phrase "best effort"?
> Or the notion that IP doesn't guarantee preserving
> packet sequence, or even guarantee delivery?

Specifically, "best effort" because that is not the same thing as
unreliable or out-of-order delivery (clearly mentioned in Pouzin's
early papers) or type-of-service networking (also in Pouzin and RFC791).

To be clear, packets delivered by a "best effort" mechanism and
packets delivered by a "type of service" or "differentiated services"
mechanism are all subject to loss or out-of-order delivery and therefore
all trigger the end to end argument.

As I said, Internet governance people mix up "best effort" with the
end to end argument, but the latter is concerned with *unreliable*
delivery, not with best effort specifically.

By the 1990s it was a commonplace that Internet service was best effort,
but that was by contrast with the unused TOS feature of RFC791.

    Brian

> 
> I assume not the latter, because the Catenet paper said:
> 
> "It is very important to note that it is explicitly assumed that
> datagrams are not necessarily kept in the same sequence on
> exiting a network as when they entered.  Furthermore, it is
> assumed that datagrams may be lost or even duplicated within the
> network.  It is left up to higher level protocols in the catenet
> model to recover from any problems these assumptions may
> introduce."
> 
> And RFC791 restates this as:
> 
> "The internet protocol does not provide a reliable communication
> facility.  There are no acknowledgments either end-to-end or
> hop-by-hop.  There is no error control for data, only a header
> checksum.  There are no retransmissions.  There is no flow control."
> 
> To my mind this explains what was later more loosely referred to
> as "best effort".  Is there any reason to believe it was used as
> a term of art, rather than a shorthand description of the 2 paragraphs
> I quoted? (I know that when I first heard "best effort" I
> interpreted it as a summary, but that just describes my reaction,
> not the intent of whoever used it).
> 
> On 2017-01-19 11:09, Guy Almes wrote:
>> Hi Brian,
>>    This is actually an interesting question.
>>    During the 1990s we had a discussion in which this phrase came up,
>> used in the now-typical networking sense.
>>    One colleague in the conversation, a lawyer, who had done work in 
>> the
>> construction industry, found the usage very odd because, he said, in
>> contracts within that industry the phrase had a specific meaning and it
>> obliged a person/company to a very very high standard of "best effort".
>> In that context, for example, it might oblige a company to spend
>> money/effort to a degree that would keep a promise but ruin any hope 
>> for
>> making money in the deal.  He noted that we computer engineers were
>> using the term in an almost opposite (and, to him, an ironic) sense,
>> viz., "do whatever is normal, but *not* heroic, and if it works, good
>> and if it doesn't, don't worry about it".
>>
>>    If the phrase is connected that older "contract language" usage, it
>> would indeed be interesting to see how it came to have almost a 
>> reversal
>> or sense.
>>
>> 	-- Guy
>>
>> On 1/18/17 2:34 PM, Brian E Carpenter wrote:
>>> I learnt very early on that the Internet offered a "best effort" 
>>> service
>>> for the delivery of datagrams.
>>>
>>> Where did that meme come from, and when?
>>>
>>> The earliest trace I found in a quick trawl was 1986 (RFC992). But 
>>> RFC791
>>> doesn't mention it, and defined TOS, such that all packets were *not* 
>>> assumed
>>> to be created equal. The 1984 Saltzer et al paper doesn't mention it 
>>> either.
>>>
>>> (RFC768 does say that UDP delivery is "not guaranteed" but that is not
>>> the same thing as "best effort".)
>>>
>>> The question is of interest because some analyses of network 
>>> neutrality,
>>> including a student dissertation I was reviewing yesterday, conflate 
>>> the
>>> end-to-end principle with best-effort packet delivery.
>>>
>>> Regards
>>>      Brian
>>> _______
>>> internet-history mailing list
>>> internet-history at postel.org
>>> http://mailman.postel.org/mailman/listinfo/internet-history
>>> Contact list-owner at postel.org for assistance.
>>>
>> _______
>> internet-history mailing list
>> internet-history at postel.org
>> http://mailman.postel.org/mailman/listinfo/internet-history
>> Contact list-owner at postel.org for assistance.
> 


From jeanjour at comcast.net  Fri Jan 20 12:10:39 2017
From: jeanjour at comcast.net (John Day)
Date: Fri, 20 Jan 2017 15:10:39 -0500
Subject: [ih] Autodin II
In-Reply-To: <8F90431E-E2C2-4E99-A684-4A348C2E1A58@planet.nl>
References: <8F90431E-E2C2-4E99-A684-4A348C2E1A58@planet.nl>
Message-ID: <DF11ADCA-A966-4ECA-B8BD-42D2C60AB87C@comcast.net>

AUTODIN II was the typical big DOD project of the 70s that made beltway bandits like CSC, Booz-Allen, SDC, etc. very happy. It was suppose to be a multi-level secure network. Several teams bid on the network. (You will laugh). The bid was won by Western Union, CSC, and Aero Ford. On the day it was awarded I predicted it would fail. It quickly became a morass of cost overruns and overly complex specs. At one point we pointed out to the Contracting Officer that it would be cheaper to build a network for each security domain than building AUTODIN II.  By that time, it was clearly dead in the water and dying. They did it a favor killing it.

John

> On Jan 20, 2017, at 10:01, Paul Ruizendaal <pnr at planet.nl> wrote:
> 
> 
> I'm currently retyping the Wingfield 1978/79 TCP/IP implementation (written for Unix V6/V7, in C); the relevant spec appears to be IEN54 & 55
> 
> This TCP/IP stack has a significant portion of code dealing with security, precedence and user groups (TCC's). From the code it seems that this is related to facilities offered by the AUTODIN II network. IEN55 mentions in its history section that some work on TCP2 was done for AUTODIN II in 1976.
> 
> A quick glance at Google results suggests that AUTODIN II was a failing project, cancelled in 1982. However, it must have been seen as viable and relevant in the '76-'79 timeframe.
> 
> What is the connection between TCP/IP work in the 70's and Autodin II? If it has been discussed before, I would much appreciate a link.
> 
> Thanks,
> 
> Paul
> 
> 
> 
> 
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.




From bill.n1vux at gmail.com  Fri Jan 20 13:26:58 2017
From: bill.n1vux at gmail.com (Bill Ricker)
Date: Fri, 20 Jan 2017 16:26:58 -0500
Subject: [ih] Autodin II
In-Reply-To: <DF11ADCA-A966-4ECA-B8BD-42D2C60AB87C@comcast.net>
References: <8F90431E-E2C2-4E99-A684-4A348C2E1A58@planet.nl>
	<DF11ADCA-A966-4ECA-B8BD-42D2C60AB87C@comcast.net>
Message-ID: <CAAbKA3W2mAudTFNErL45cXLp7X=MAkmgiCCAQVB-ONARhXQrDg@mail.gmail.com>

On Fri, Jan 20, 2017 at 3:10 PM, John Day <jeanjour at comcast.net> wrote:
> It quickly became a morass of cost overruns and overly complex specs. At one point we pointed out to the Contracting Officer that it would be cheaper to build a network for each security domain than building AUTODIN II.

At which point, the priority of certified, semi-automated
human-in-loop down-grade-review "Guard" mini-computer systems to
bridge the air-gaps betwixt the multiple single-level domains became
more urgent.


From vint at google.com  Fri Jan 20 16:06:13 2017
From: vint at google.com (Vint Cerf)
Date: Fri, 20 Jan 2017 19:06:13 -0500
Subject: [ih] Autodin II
In-Reply-To: <8F90431E-E2C2-4E99-A684-4A348C2E1A58@planet.nl>
References: <8F90431E-E2C2-4E99-A684-4A348C2E1A58@planet.nl>
Message-ID: <CAHxHggcNtSG-pqf2DU-2WN14YpmUhM-x4XR7kKxKebmP6xsZXA@mail.gmail.com>

Autodin II took the December 1974 TCP spec (RFC 675) and elaborated on that
in a fork independent from the evolution of the RFC series TCP and TCP/IP.
The fork was a huge pain for me during that time. The project was top heavy
and eventually crashed and burned. The best source of information for that
story is Peter Sevcik who now runs NetForecast, Inc - in Charlottesville if
I remember correctly.

v


On Fri, Jan 20, 2017 at 10:01 AM, Paul Ruizendaal <pnr at planet.nl> wrote:

>
> I'm currently retyping the Wingfield 1978/79 TCP/IP implementation
> (written for Unix V6/V7, in C); the relevant spec appears to be IEN54 & 55
>
> This TCP/IP stack has a significant portion of code dealing with security,
> precedence and user groups (TCC's). From the code it seems that this is
> related to facilities offered by the AUTODIN II network. IEN55 mentions in
> its history section that some work on TCP2 was done for AUTODIN II in 1976.
>
> A quick glance at Google results suggests that AUTODIN II was a failing
> project, cancelled in 1982. However, it must have been seen as viable and
> relevant in the '76-'79 timeframe.
>
> What is the connection between TCP/IP work in the 70's and Autodin II? If
> it has been discussed before, I would much appreciate a link.
>
> Thanks,
>
> Paul
>
>
>
>
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.
>



-- 
New postal address:
Google
1875 Explorer Street, 10th Floor
Reston, VA 20190
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170120/98c2b807/attachment.htm>

From pnr at planet.nl  Fri Jan 20 18:31:17 2017
From: pnr at planet.nl (Paul Ruizendaal)
Date: Sat, 21 Jan 2017 03:31:17 +0100
Subject: [ih] Autodin II
Message-ID: <0A254823-337E-4D71-8854-45828B5F2C97@planet.nl>

Many thanks for those clarifications!

The driver code for Autodin II indeed appears significantly more complex than that for Arpanet.

So in late '78, early '79 did Autodin II exist as a network, or was it just a bunch of specs?

And - at that time - was the TCP in use on / proposed for Autodin II the same as that in IEN55?

Paul




From mfidelman at meetinghouse.net  Fri Jan 20 19:13:40 2017
From: mfidelman at meetinghouse.net (Miles Fidelman)
Date: Fri, 20 Jan 2017 22:13:40 -0500
Subject: [ih] Autodin II
In-Reply-To: <0A254823-337E-4D71-8854-45828B5F2C97@planet.nl>
References: <0A254823-337E-4D71-8854-45828B5F2C97@planet.nl>
Message-ID: <20c91ce8-8c5a-a7e6-024b-c42cc427e795@meetinghouse.net>



On 1/20/17 9:31 PM, Paul Ruizendaal wrote:
> Many thanks for those clarifications!
>
> The driver code for Autodin II indeed appears significantly more complex than that for Arpanet.
>
> So in late '78, early '79 did Autodin II exist as a network, or was it just a bunch of specs?

It was a procurement.  It never actually got built as a system.  The 
original AUTODIN message system kept getting updated incrementally, and 
traffic kept moving onto email.  There were still pieces of the service 
still in operation as of a few years ago - there still well could be.


>
> And - at that time - was the TCP in use on / proposed for Autodin II the same as that in IEN55?
>
> Paul
>
>
> _______
> internet-history mailing list
> internet-history at postel.org
> http://mailman.postel.org/mailman/listinfo/internet-history
> Contact list-owner at postel.org for assistance.

-- 
In theory, there is no difference between theory and practice.
In practice, there is.  .... Yogi Berra



From amckenzie3 at yahoo.com  Sun Jan 22 09:50:12 2017
From: amckenzie3 at yahoo.com (Alex McKenzie)
Date: Sun, 22 Jan 2017 17:50:12 +0000 (UTC)
Subject: [ih] Autodin II
References: <1026980157.2116514.1485107412169.ref@mail.yahoo.com>
Message-ID: <1026980157.2116514.1485107412169@mail.yahoo.com>

Friends,
For some reason my responses to postings in Internet History seem to be falling into a black hole.
There is an article titled "ARPANET, the Defense Data Network, and Internet" written by Dave Walden and me, which appeared in Encyclopedia of Telecommunications - Volume 1, edited by Froehlich and Kent.? It has a bit to say about the decision to replace the Autodin II technology by ARPANET technology.
Cheers,Alex McKenzie



?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170122/8c675424/attachment.htm>

From dave.walden.family at gmail.com  Sun Jan 22 10:06:12 2017
From: dave.walden.family at gmail.com (dave.walden.family at gmail.com)
Date: Sun, 22 Jan 2017 13:06:12 -0500
Subject: [ih] Autodin II
In-Reply-To: <1026980157.2116514.1485107412169@mail.yahoo.com>
References: <1026980157.2116514.1485107412169.ref@mail.yahoo.com>
	<1026980157.2116514.1485107412169@mail.yahoo.com>
Message-ID: <6CE7A57B-7C2D-4CAC-B5E5-3ED18EB6CBF7@gmail.com>

A preprint of the encyclopedia article is at
     http://walden-family.com/public/encyclopedia-article.pdf       

Sent from my iPad

On Jan 22, 2017, at 12:50 PM, Alex McKenzie <amckenzie3 at yahoo.com> wrote:

> Friends,
> 
> For some reason my responses to postings in Internet History seem to be falling into a black hole.
> 
> There is an article titled "ARPANET, the Defense Data Network, and Internet" written by Dave Walden and me, which appeared in Encyclopedia of Telecommunications - Volume 1, edited by Froehlich and Kent.  It has a bit to say about the decision to replace the Autodin II technology by ARPANET technology.
> 
> Cheers,
> Alex McKenzie
> 
> 
> 
>  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://elists.isoc.org/pipermail/internet-history/attachments/20170122/5e9c0b1c/attachment.htm>

From lists at eitanadler.com  Sun Jan 22 12:39:48 2017
From: lists at eitanadler.com (Eitan Adler)
Date: Sun, 22 Jan 2017 12:39:48 -0800
Subject: [ih] Autodin II
In-Reply-To: <1026980157.2116514.1485107412169@mail.yahoo.com>
References: <1026980157.2116514.1485107412169.ref@mail.yahoo.com>
	<1026980157.2116514.1485107412169@mail.yahoo.com>
Message-ID: <CAF6rxg=DXdR7j9N6oxmoA97km02VZyfCSs+Zn+ubX2ZcZ=J9sw@mail.gmail.com>

On 22 January 2017 at 09:50, Alex McKenzie <amckenzie3 at yahoo.com> wrote:
> Friends,
>
> For some reason my responses to postings in Internet History seem to be
> falling into a black hole.

This might be cause yahoo.com is mailing list hostile in that it does
not allow non-yahoo addresses to send mail on your behalf. This
results in emails failing DMARC tests.

Authentication-Results: mx.google.com;
       dkim=neutral (body hash did not verify) header.i=@yahoo.com;
       spf=neutral (google.com: 128.9.160.161 is neither permitted nor
denied by best guess record for domain of
internet-history-bounces at postel.org)
smtp.mailfrom=internet-history-bounces at postel.org;
       dmarc=fail (p=REJECT sp=REJECT dis=NONE) header.from=yahoo.com



-- 
Eitan Adler


From dhc2 at dcrocker.net  Sun Jan 22 15:01:02 2017
From: dhc2 at dcrocker.net (Dave Crocker)
Date: Sun, 22 Jan 2017 15:01:02 -0800
Subject: [ih] Autodin II
In-Reply-To: <CAF6rxg=DXdR7j9N6oxmoA97km02VZyfCSs+Zn+ubX2ZcZ=J9sw@mail.gmail.com>
References: <1026980157.2116514.1485107412169.ref@mail.yahoo.com>
	<1026980157.2116514.1485107412169@mail.yahoo.com>
	<CAF6rxg=DXdR7j9N6oxmoA97km02VZyfCSs+Zn+ubX2ZcZ=J9sw@mail.gmail.com>
Message-ID: <c8f863dd-2646-4489-abf5-47f2662c5f53@dcrocker.net>

As the auth-results header field shows, the specific issue is their 
publishing a DMARC record directing the receiving host to reject mail 
that does not validate DKIM (using the author From field domain) or SPF. 
  Their use of DMARC this was set new precedent and it is indeed mailing 
list unfriendly.  There currently is no fix for it from the DMARC publisher.

The only hack that seems to work is modifying the author From: field 
address to use a different domain.  This often conforms to a template 
such as:

    From:  Original Display Name via listname
           <listname at listhost.example.>

or the like.

The premise that this somehow provides generally-useful anti-abuse 
protection is, shall we say, controversial.  However it does preclude 
unauthorized third-party's use of the originating host's domain name in 
the From: field...

d/

On 1/22/2017 12:39 PM, Eitan Adler wrote:
> On 22 January 2017 at 09:50, Alex McKenzie <amckenzie3 at yahoo.com> wrote:
>> Friends,
>>
>> For some reason my responses to postings in Internet History seem to be
>> falling into a black hole.
>
> This might be cause yahoo.com is mailing list hostile in that it does
> not allow non-yahoo addresses to send mail on your behalf. This
> results in emails failing DMARC tests.
>
> Authentication-Results: mx.google.com;
>        dkim=neutral (body hash did not verify) header.i=@yahoo.com;
>        spf=neutral (google.com: 128.9.160.161 is neither permitted nor
> denied by best guess record for domain of
> internet-history-bounces at postel.org)
> smtp.mailfrom=internet-history-bounces at postel.org;
>        dmarc=fail (p=REJECT sp=REJECT dis=NONE) header.from=yahoo.com
>
>
>


-- 

   Dave Crocker
   Brandenburg InternetWorking
   bbiw.net