[ih] capacity v bandwidth

Sun May 31 12:27:55 PDT 2026

Jack,

That's delightful.  Reminds me of hearing about a lengthy discussion
between, IIRC, Corby and Weizenbaum about 645.  Apparently, the discussion
went on for a while before they realized one person was talking about the
course and the other about the machine.

Collision among TLAs   is, of course, rampant.  I switched to four letter
acronyms many years ago.  You're probably familiar with the birthday
problem: how big does a class have to be before there's a 50% chance that
at least two students will have the same birthday?  The answer is just 23.
For TLA, assuming a uniformly random distribution, the 50% level is just
155 TLAs.  And the distribution is not uniform at all, so the 50% level is
fewer than 155.

Steve

On Sun, May 31, 2026 at 3:13 PM Jack Haverty via Internet-history <
internet-history at elists.isoc.org> wrote:

> Turning purple is a clear sign of psychological stress.   Vint's escort
> was probably contemplating the massive amount of paperwork required to
> document the incident, the disciplinary hearings that would follow, as
> well as considering what his or her new life would be like at the new
> assignment on one of the outer islands of the Aleutians.
>
> I can only recall small bits of my own meeting experience.  I'm not even
> sure of where I was working at the time.   It may have been while I was
> at BBN, and involving a meeting somewhere in the metro DC area.  Or it
> may have been when I had a student job at the MIT Instrumentation Lab,
> where navigation systems were developed for use in Apollo and
> Minuteman.   Anyway, it was a long time ago, in the late 1960s or 1970s.
>
> There was a meeting, and I was late for it.   I wasn't a participant
> giving a presentation, but more like a fly on the wall to learn about
> some project.  After getting through the check-in procedure in the
> Entrance (with security clearances it could take a while...), I went
> down a corridor, saw a meeting already in progress with a few dozen
> people, most of them around a large table engaged in intense debate.  I
> found a chair against the wall and quietly sat down.  No one noticed.
>
> The discussion involved DIA and DCA, which apparently "don't work well
> together".  As a techie nerd, I had heard of DIA and DCA, but hadn't had
> any actual experience with them.  They were both part of IBM's technical
> architecture.   DIA was Document Interchange Architecture, and DCA was
> Document Content Architecture.  I could believe that they didn't work
> well together -- possibly designed by different committees or groups
> within IBM.
>
> The discussion seemed to be focussed on comparing different scenarios,
> and predicting what would happen in each.  The main difference between
> scenarios was apparently associated with "new debts".  I thought perhaps
> the current topic had something to do with budgeting.  But someone
> mentioned "new debts per second", which seemed a little strange.  I
> could believe "per year" or "per quarter" would make sense for financial
> planning and decisions.  But "per second"!?
>
> Finally someone put up a slide or document for some scenario.   I had
> heard what I expected to hear - "new debts".  But the slide revealed it
> was actually "nudets".  I realized the discussion had been about NUclear
> DETonations in different scenarios.  The "per second" part was
> frightening, which is probably why I still remember it.
>
> I was clearly in the wrong meeting.  Here, DIA was Defense Intelligence
> Agency.  DCA was Defense Communications Agency. Nothing to do with IBM
> technology.
>
> I quietly got up and left the room.
>
> I learned that day that the world suffers from a serious shortage of
> acronyms.  IP addresses are easy, just make the numbers longer.  But
> TLAs have to be just three letters, and that's not enough to avoid
> conflicts and ambiguity.
>
> /Jack Haverty
>
>
> On 5/31/26 10:59, the keyboard of geoff goodfellow via Internet-history
> wrote:
> > thanks vint! AND believe it or not (now that we know what the room was)
> it
> > just so happens there is a "vintage" 1976(!) picture of the NMCC at
> > https://en.wikipedia.org/wiki/National_Military_Command_Center
> >
> > yours truly is now wondering what a present day/"modern era" of it must
> > look like... :)
> >
> > g
> >
> > On Sun, May 31, 2026 at 10:47 AM Vint Cerf <vint at google.com> wrote:
> >
> >> Thanks for asking, Geoff.
> >>
> >> On my first day of work at ARPA (as it was then called in 1976), I had
> to
> >> report to the Pentagon to pick up my DOD building pass. I went into the
> >> River Entrance and the guard told me to take a seat. Not long after, a
> guy
> >> came out of the bowels of the Pentagon, conferred with the guard who
> >> pointed at me. The man came over to me and said something that I didn't
> >> really hear [you all know I wear hearing aids ...] but assumed it was
> >> something like "good morning" - so I nodded and he escorted me into one
> of
> >> the corridors. Soon we encountered another guard (Marine, armed) and
> they
> >> attached another badge. We continued deeper into the Pentagon, I picked
> up
> >> additional badges on the way until we entered a very big space with a
> lot
> >> of people with headsets, large displays, even larger displays on the
> walls.
> >> I said "This looks pretty elaborate for a personnel office." My escort
> >> looked startled, turned purple (no kidding!)
> >>
> >> "Aren't you Mr. Smith!!??" he said. "No, I'm Vint Cerf. I'm just here to
> >> pick up my building pass."  My escort fell silent, whipped me around,
> said
> >> not another word, marched me through the reverse path, shedding badges
> as
> >> we went. I ended up back in the River Entrance. I had just inadvertently
> >> visited the National Military Command Center. In all the years since,
> >> despite my clearances, I have never returned to the NMCC!
> >>
> >> Vint Cerf
> >>
> >>
> >> On Sun, May 31, 2026 at 12:53 PM the keyboard of geoff goodfellow via
> >> Internet-history <internet-history at elists.isoc.org> wrote:
> >>
> >>> vint, any chance you would be willing to recount to the list what
> happened
> >>> at the Pentagon on your first day at ARPA in 1976.. in how you
> mistakingly
> >>> got mistook for someone else in the reception area.. and ended up being
> >>> walked thru multiple levels of guarded "security" checkpoints and
> ended up
> >>> in the "War Room" (or whatever it was) instead of the HR office
> (recalling
> >>> that ARPA didn't have an HR office)?
> >>>
> >>> over the years yours truly has retold that story of yours -- that iirc
> you
> >>> to related to us at a meeting or conference or something -- when people
> >>> have said such-and-such a "security" breach thing was "fake"/"made up"
> and
> >>> couldn't ever possibly have happened in real life (such as say the
> >>> security
> >>> breach incident in the 1997 movie "Contact" that resulted in deliberate
> >>> sabotage) where someone "got/slipped through security"
> >>>
> >>> g
> >>>
> >>> On Sun, May 31, 2026 at 9:07 AM vinton cerf via Internet-history <
> >>> internet-history at elists.isoc.org> wrote:
> >>>
> >>>> Yes, I left Stanford in July or August and started at ARPA in Sept
> 1976.
> >>>>
> >>>> V
> >>>>
> >>>> On Sun, May 31, 2026, 11:39 Lawrence Stewart via Internet-history <
> >>>> internet-history at elists.isoc.org> wrote:
> >>>>
> >>>>> … Rummages around for information theory hat.
> >>>>>
> >>>>> I don’t think I am alone in holding that things with digital
> >>> interfaces
> >>>>> should say “bandwidth” and things with analog interfaces should say
> >>>>> “capacity”.
> >>>>>
> >>>>> The Shannon-Hartley theorem is about the analog part of this.
> >>>>>
> >>>>> In the digital domain, I think almost everyone uses “bandwidth” to
> >>> talk
> >>>>> about peak available bits-per-second and “capacity” to talk about
> >>> usable
> >>>>> bits-per-second under certain loads, queuing policies, and the like.
> >>>>>
> >>>>> In the above, bandwidth is absolutely misused for digital circuits as
> >>>>> compared to its analog meaning.  Capacity, unfortunately, has two
> >>>> meanings
> >>>>> and they get muddled all the time.
> >>>>>
> >>>>> Language evolves and non of this offends me.
> >>>>>
> >>>>> Regarding Vint at Stanford I arrived there in September 1976, and by
> >>> the
> >>>>> time I was corresponding with Vint about using Ron Crane’s VDH line
> >>>> driver
> >>>>> to get SU-ISL connected to the SUMEX IMP, he was back in Washington I
> >>>> think.
> >>>>> -L
> >>>>>
> >>>>>
> >>>>> --
> >>>>> Internet-history mailing list
> >>>>> Internet-history at elists.isoc.org
> >>>>> https://elists.isoc.org/mailman/listinfo/internet-history
> >>>>> -
> >>>>> Unsubscribe:
> >>>>>
> >>>
> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history
> >>>> --
> >>>> Internet-history mailing list
> >>>> Internet-history at elists.isoc.org
> >>>> https://elists.isoc.org/mailman/listinfo/internet-history
> >>>> -
> >>>> Unsubscribe:
> >>>>
> >>>
> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history
> >>>>
> >>> --
> >>> 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
> >>> -
> >>> Unsubscribe:
> >>>
> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history
> >>>
> >>
> >> --
> >> After July 1, Please send any postal/overnight deliveries to:
> >> Vint Cerf
> >> 1435 Woodhurst Blvd
> >> McLean, VA 22102
> >>
> >> Until July 1, send postal/overnight deliveries to:
> >> Google, LLC
> >> 1900 Reston Metro Plaza, 16th Floor
> >> Reston, VA 20190
> >> +1 650-224-2788
> >>
> >>
> >>
> >>
> >>
> >>
> >>
>
> --
> Internet-history mailing list
> Internet-history at elists.isoc.org
> https://elists.isoc.org/mailman/listinfo/internet-history
> -
> Unsubscribe:
> https://app.smartsheet.com/b/form/9b6ef0621638436ab0a9b23cb0668b0b?The%20list%20to%20be%20unsubscribed%20from=Internet-history
>

-- 
Sent by a Verified

sender