[ih] Larry Roberts & RD the first electronic mail manager software [was written in TECO on TENEX] (fwd)

Tue Aug 8 17:30:23 PDT 2023

for some unknown reason the IH list is randomly not sending out items sent
to it... luckily, Steve sent yours truly this reply directly -- now being
fwd'd:

---------- Forwarded message ---------
From: Steve Crocker <steve at shinkuro.com>
Date: Tue, Aug 8, 2023 at 5:23 PM
Subject: Re: [ih] Larry Roberts & RD the first electronic mail manager
software [was written in TECO on TENEX]
To: the keyboard of geoff goodfellow <geoff at iconia.com>
Cc: Internet-history <internet-history at elists.isoc.org>

On Tue, Aug 8, 2023 at 12:52 PM the keyboard of geoff goodfellow <
geoff at iconia.com> wrote:

> that was a really excellent "little bit" steve... if yours truly may add a
> little bit more to "I'm interrupted by a chat request from Lukasik"
> vis-a-vis the "chat request":
>
> that was known as the LINK command in the Tenex EXEC, where you could say
> LINK USERNAME or link TTY# and your terminal session and the other person's
> terminal session would be summarily LINKED to each other -- where each
> would see exactly what the other was seeing... IIRC up to 5 terminals could
> be LINKED together... ERGO, after Lukasik LINKED to you, he likely just
> typed out the memo file on his terminal which of course came out on yours.
>

Yes, that sounds right.  I don't remember typing the semicolon, but it was
obviously necessary.

An earlier hack, not involving Lukasik or Roberts was to login to a Tenex
machine, Telnet back to the same machine, and then link the two sessions.
Typing a single character caused an endless flow.  A related but different
issue arose when typing control-C in a session that was a Telnet session to
another machine.  Would the control-C be interpreted by the first session
or the second session?  Some wag coined the term cybergammaphobia,
inordinate fear of control-C.

Thanks,

Steve

>
> when terminals were linked to one another, we would proceed our text with
> an ";" -- which was the EXEC's "comment" character, i.e.
> @; hi there
> @; what's up?
> [...]
> with the occasional "flub" of where one would forget to proceed their
> comment with the ";" and the EXEC would then interpret it as commands...
> with sometimes "unusual" results... later, a program named TALK was written
> what just took what was typed into it and did nothing with it... although
> it did have a nice feature in that you could type stuff in without RETURNs
> and it would automatically insert line breaks (CR's) after 60 some odd
> characters or so... recalling that in those days we mostly had Teletype
> Model 33 ASR's (or KSR's) that were only upper case and had a line length
> of 72 characters (and printed out at 110 baud, or 10 characters per second
> CPS)
>
> one last thing about LINKing on Tenex: if you were a WHEEL (I.e.
> privileged user) it was possible to do one-way LINKS -- i.e. SPY on a
> terminal or user... this was most useful for monitoring the CTY and/or TTY0
> which is where the monitor (or system daemon's/sysjob's) would spit out
> "issues".
>
> geoff
>
> On Tue, Aug 8, 2023 at 9:24 AM Steve Crocker <steve at shinkuro.com> wrote:
>
>> Adding a little bit to your summary:
>>
>> I was in (D)ARPA/IPTO from mid 1971 to mid 1974.  When I arrived, the
>> Arpanet had been in operation for almost two years.  IPTO consisted of the
>> director (Larry Roberts), three program managers (Bruce Dolan, John Perry,
>> me), a senior non-technical person who handled our budgets and
>> related matters (Al Blue), and two secretaries.  Steve Lukasik was the
>> (D)ARPA director and hence Larry's boss.  I write (D)ARPA here because the
>> agency was christened the Advanced Research Projects Agency (ARPA) and
>> embedded within the Office of the Secretary of Defense (OSD) in 1958.  On 1
>> July 1972 it was moved out of OSD and became a separate Defense Agency.  In
>> the process, it acquitted the initial D, i.e. Defense Advanced Research
>> Projects Agency, hence DARPA.  The transition made essentially no
>> difference in our mission or operation.
>>
>> A short time later, we had a TIP installed in the office.  This meant we
>> could connect  a large number of terminals.  Everyone in our office had a
>> terminal on the desk, and soon others through the agency had terminals
>> too.  Lukasik was a strong proponent and soon required that each of his
>> direct reports, i.e. the directors of the other Offices, use email to
>> communicate with him. Each of us was also given an account on the Tenex
>> machine at the new USC Information Sciences Institute (USC-ISI).
>>
>> Lukasik used email to transform the way he conducted business.  Even
>> though we were all housed in the same office building -- 1400 Wilson Blvd,
>> Arlington, VA, where a plaque now stands commemorating the creation of the
>> Arpanet -- and could meet face to face easily, he found email made it
>> possible for him to reorganize his time and attention.  (More on this
>> below.)
>>
>> The earliest made reader we had was RD.  I believe Roberts wrote.  He was
>> a very competent TECO hacker, and he whipped it up quickly.  In that
>> version, mail was stored in a continuous file.  In some later mail systems,
>> each message was in a separate file, but at the beginning all of a user's
>> mail was in a continuous file.  These files grew fairly quickly.  Moving to
>> a particular message became a slow process.
>>
>> I too was a fairly competent TECO hacker.  I took a look and realized why
>> it was taking too long to find a message.  In the file, each message was
>> preceded by a line that had a string indicating it was the beginning of a
>> message and a count of the number of characters in the message.  In
>> principle, that should have made it quick to move forward in the file from
>> one message to the next.  However, that count didn't coincide with the
>> actual number of characters in the file.  When the count was created, it
>> treated the end of each line as one character, the newline character (NL),
>> but when the message arrived, it became two characters, carriage return
>> (CR) followed by line feed (LF).  Hence the TECO code had a loop that moved
>> forward one character at a time until it found the beginning of the next
>> message.  Very slow.  (I might be misremembering and reversing these
>> effects, but either way it was the discrepancy between the way ends of
>> lines were demarcated at the point of creation and the the way ends of
>> lines were demarcated on the receiving side.)
>>
>> I replaced that loop with code that skipped forward a line at a time and
>> then adjusted the number of characters it still had to go.  The speedup was
>> dramatic.
>>
>> Above I alluded to changes in Lukasik's management style.  My primary
>> focus while I was at (D)ARPA was the AI portfolio, i.e., MIT, CMU,
>> Stanford, SRI and BN and others.  AI was unquestionably a long term bet.
>> (D)ARPA's usual profile was heavy funding of some area for five years or so
>> to trigger a large change.  ("A factor of ten, not ten percent" was the
>> common catchphrase.). We knew AI was a *much* longer trek.  Fifty years,
>> most likely.  That meant it was a constant challenge to protect the funding
>> and explain the research to Congress.
>>
>> In the UK, Science Research Council, the UK counterpart to the US NSF,
>> had been funding Donald Michie's AI research at the University of
>> Edinburgh.  According to a 29 June 1973 news article in Science, "In early
>> 1972 Sir James Lighthill of Cambridge University undertook to survey the
>> field of artificial intelligence (AI) for the Science Research Council of
>> Britain."  His report was dismissive of AI and had a strongly negative
>> influence on AI funding in the UK.  Toward the end of the Science news
>> article, the following was embedded:
>>
>> Even granting that AI is an intellectually important area for research,
>> it is fair to ask whether the field is using its resources wisely. The
>> Lighthill report suggests that, in the United States especially, little
>> attention has been given to this question, in part because there has been a
>> relatively assured source of funding.  As is true for computer science
>> in general, research on AI is pre-dominantly supported by the Defense
>> Advanced Research Projects Agency (ARPA), which provides about $4.5 million
>> a year.
>>
>> 29 July 1973 was a Friday.  That evening I was sitting at my terminal at
>> home.  Abruptly, I'm interrupted by a chat request from Lukasik.  I had no
>> idea he even knew how to do that.  After we were connected, he asked if he
>> printed something on his terminal, would I see it?  I assured him I would.
>> Immediately a ten point memo starts printing on my terminal.  He's
>> concerned that Congress might see this comment in Science and take aim at
>> our AI program.  His memo dealt with possible press or Congressional
>> questions, coordination internally, etc.  I understood completely he
>> perceived a possible threat and was making sure we were manning the
>> ramparts.  Fortunately, the attack never came.  Even so, we did have a
>> session in the office on a Sunday afternoon later in the summer watching a
>> taped copy of the debate between McCarthy and Lighthill.
>>
>> AI wars aside, the relevance to this thread is how quickly and
>> effectively Lukasik absorbed the new technology and put it to use.
>>
>> Steve
>>
>>
>>
>> On Mon, Aug 7, 2023 at 8:48 PM the keyboard of geoff goodfellow via
>> Internet-history <internet-history at elists.isoc.org> wrote:
>>
>>> "... In 1969 Roberts became director of the Information Processing
>>> Techniques Office at ARPA. In 1971 he wrote one of the first e-mail
>>> programs, RD, which for the first time allowed users to save, delete, and
>>> organize their messages....
>>> so sez:
>>> https://www.britannica.com/biography/Lawrence-Roberts
>>>
>>> while
>>>
>>> https://cs.stanford.edu/people/eroberts/courses/soco/projects/1999-00/internet/email.html
>>>  sez:
>>>
>>> ...The Rise of Email by way of Convenient Mail-Managers
>>>
>>> Stephen Lukasik, one of ARPA's directors from 1971-75 might be one of the
>>> most important advocates of email in its young age. He strongly
>>> encouraged
>>> his staff to use email. It was even considered the most effective way to
>>> reach Lukasik. Close friends and colleagues with Larry Roberts, Lukasik
>>> confided to Roberts in 1973 that he was beginning to face problems
>>> organizing the chaotic pile of messages received each day. (At this
>>> point,
>>> in a study conducted by ARPA under Lukasik's direction discovered that
>>> 75%
>>> of all ARPANET traffic was in the form of email.) Not to mention, reading
>>> and responding to the messages was, as of yet not a simple matter. In
>>> response, Roberts wrote the first "mail-manager software," RD. Its
>>> functions included displaying a menu of messages and allowing the user to
>>> file messages and delete messages. This inspired an outburst of
>>> variations
>>> on the mail manager. A legendary variation, MSG was created in 1975 by
>>> John
>>> Vittal. MSG quickly became the most widely used mail-management program
>>> due
>>> to its convenient features. It could handle an even greater amount of
>>> mail
>>> than RD, could sort messages into separate files, and most importantly,
>>> reply and forward with much greater ease. MSG also spawned a large amount
>>> of variations...
>>>
>>> [any others?]
>>>
>>> --
>>> Geoff.Goodfellow at iconia.com
>>> living as The Truth is True
>>> --
>>> Internet-history mailing list
>>> Internet-history at elists.isoc.org
>>> https://elists.isoc.org/mailman/listinfo/internet-history
>>>
>>
>
> --
> Geoff.Goodfellow at iconia.com
> living as The Truth is True
>
>

-- 
Geoff.Goodfellow at iconia.com
living as The Truth is True