My first two years at BBN

I moved to BBN in September 1967.  It was very different than MIT
Lincoln Laboratory from which I had come, and not only in that I had my
own interactive computer terminal in my office in BBN's 20 Moulton
Street building (an our warehouse). The other staff members were closer
in (young) age to me than my colleagues at Lincoln Lab had been.  Many
of us were primarily computer programmers rather than physicists,
mathematicians or electrical engineers first and programmers second as
had been the case at Lincoln Lab.  Things were less formal at BBN.
Staff members had stopped bringing their dogs to work before I got
there, but Bernie Cosell and Andy Munster were there playing their
musical instruments during work hours, and Bob Brooks was around and
about wearing no shoes.  One of the first computer oriented things I
saw at BBN was the output of Cosell's Snow program which printed
banners (such as "Happy Birthday") on a line printer.

There was a refrigerator at the end of the hall (near our group's PDP-1
computer room), and I drank gallons of Coke a day from the refrigerator
on some sort of honor payment system.  Cosell recalls, "This was a hack
by me and Andy Munster.  There used to be a regular coke machine right
across the hall from the PDP-1 computer room.  I can't remember what
they did: they probably raised the cost of it, or stopped dispensing
bottles, or who knows what ... but Andy and I were pissed and refused
to pay ... So we decided to supply Pepsi in a refrigerator -- not coke!
Andy didn't care and I was a Pepsi freak and had the swing vote :o). It
was Andy's brilliant insight that folk didn't steal because they were
poor, or for the thrill (if you made it cheap and trivial), but because
they were lazy.  So we cooked up a scheme where you could take a Pepsi
and just check the sheet on the door of the frig.  Once in a while,
Andy and I would tote up the numbers and go around and collect.  What
was amazing is that we had essentially zero shrinkage -- the scheme
really did work.

There was no cafeteria or short order grill as there had been at
Lincoln Lab, and my typical lunch was as coke and a package of three
rock hard Choco Crunch chocolate chip cookies from the vending machine.
Of course, some days a group of us would  go out the back door of 20
Moulton, across the alley, and into Joyce Chen's legendary Chinese
restaurant where the lunch time special was available for very little
money.  (Joyce Chen shared her building with a meat market called
Regional Beef.  Later, a Japanese restaurant named Osaka, which we
always heard was run Joyce Chen's ex-husband, replaced Joyce Chen's in
the Regional Beef building, and our easy lunch time food became a bowl
of udon noodles with some slices of meat on top.)  Eventually, Frank
made a deal with a local sandwich shop to sell their sandwiches at BBN
during lunch hour, and I took to eating these.

There was a lot of game playing going on, often at lunch hour in Mike
Fein's office.  Klab was a popular card game for a while, and Contract
Bridge was a continuing activity -- I pretty much avoided the later
because I wasn't much interested in lunch time hands with parallel
effort toward a bidding system with a long term partner and gradually
my interest in Bridge disappeared completely (my issue about long term
partnerships was misplaced -- this bridge game, which I always thought
of as Alex's bridge game, went on for decades (and maybe still is going
on) with very slow change in participants). In vogue for a while was
board game Twixt, a sort of Chinese checkers where one's pieces were
connect with knight's moves. This could also be played with pencils on
quadrille pad, and I remember Alex McKenzie and me playing game after
game this while while we were nominally attending a seminar on data
base management at the Hotel Somerset in Boston's Back Bay
neighborhood. There was also a room with the division dart board that
housed pretty regular darts matches.

Frank Heart's Computer Systems Division had several activities. The
largest was the Hospital Computer Project, led by Paul Castleman, which
was shrinking. Mike Crick (the Nobel Prize winner's son) was working
proposals or projects relating to poisen control (Bob Kahn used to drop
by our corridor to talk with Mike).  Wally Feurzeig's educational
technology group, including Paul Wexelblat, Severo Orstein, and others,
was developing LOGO turtles, among other things.  Alex McKenzie worked
on several projects or project proposals trying to get the division
into new life science activities.  Frank Heart was also getting
involved in the management of BBN's Telcomp business (Frank had
recruited Ralph Alter from Lincoln Lab to provide some technical
management to Telcomp) and that set of people were in the next corridor
of offices. And there were probably other activities I can't remember
right now.

In 1968 we began to prepare for and later bid on the the ARPANET
Request for Quotation, and I was part of this activity. However, from
when I got to BBN in 1967 until we actually began work on the ARPANET
development at the beginning of 1969, I was assigned to no long-term
project as, I assume, Frank tried to find the right long term place for
me.

I had begun graduate study in computer science at MIT in 1966 while at
Lincoln Lab and nominally continued while I was at BBN. Partly from the
courses I took there, I was very interested in computer languages and
macro processors.  In addition whatever real work I was assigned, I
also did a good bit of hacking.  I implemented Stracey's CPM macro
processor on the PDP-1 (I had previously implemented a macro processor
for an assembler at Lincoln Lab). I taught a session on compilers one
night for Frank's computer class at Northeastern's Extension Division
(or maybe it was the course of a friend of his who asked Frank to find
someone to teach this session).  At one point I taught a class or two
somehow relating to Telcomp in BBN's Program for Advanced Study, a
weekend graduate school business founded and operated for a while. I
studied TRAC in considerable depth, met Calvin Moores and Peter
Deutsch, and I think implemented a version of TRAC on the PDP-1 (Peter
had done the first, Cosell and a couple of others did another, and mine
was no earlier than the third and never used by anyone but me for
playing with TRAC).  At one point Cosell and I implemented Euler in
LISP on BBN's SDS-940; for me this was a way to study Euler and LISP.
At MIT, my never completed master's thesis work was focusing on macro
processors and extensible languages (a relatively new concept then).

As part of my infatuation with TRAC, I one day told Danny Bobrow how
wonderful I thought it was.  Perhaps because he thought I needed
something more useful to do than to revel in a dead end language like
TRAC, he got me doing reviews for ACM's Computer Surveys.  I think it
was also Danny who later got Bob Kahn and me involved with an
evaluation of SDC's XXX data management system.  This involved a trip
to an Air Force base outside of Washington, DC, to hear a seminar on
the system and trip to visit Clark Weisman at SDC in Santa Monica.

I only remember one activity involving the acoustics side of BBN. I was
assigned to take a day trip to Cincinnati with Ed Kerwin.  I don't
remember what we were there for.

During this time, the implementation of TENEX was under way at BBN.
Bobrow, Dan Murphy, Jerry Burchfiel, and Ray Tomlinson were leading
figures on this development.  I was fascinated with implementation of
virtual memories and paging (my classes at MIT were promoting the less
practical MULTICS approach) and hung around with Burchfiel, Ed Fiala,
John Barnaby, Tony Michel, etc., who were working on the project.
Barnaby later became famous for developing Word Star, which was clearly
highly influenced by Cosell's MRUNOFF which Barnaby used at BBN and
which Bernie created having been impressed Runoff at MIT.

Two projects I did during this period related to later major projects
at BBN.  In one case, I helped Alex McKenzie evaluate a bunch of
mini-computers for some life sciences proposal or project (or maybe
Alex did this alone and I just knew of what he was doing). In any case,
when it came time to choose which machine to bid for the ARPANET
project, we started with the results of the earlier evaluation project.
In other other case, I helped Paul Castleman formulate the Prophet
project.  I took trips with Paul to NIH to meet Bill Raub, we and one
of Raub's young associates visited Stanford, and in the end I wrote a
document recommending that Prophet be based on an extensible language.
To demonstrate the idea of an extensible language for Prophet, I
transcribed the Fortran code from James ???'s PhD thesis on an
extensible language into assembly language for the PDP-10. This never
really worked, and Fred Webb soon threw it away and wrote Prophet's
extensible PARSEC language from scratch.

From the beginning of 1969 (with the exception of a year in Norway from
September 1970 to September 1971), I was involved for many years with
the development of the ARPANET and then BBN's expanding computer
networking activities.  In the almost two years before I left for
Norway, the others involved with the project mostly included Will
Crowther, Bernie Cosell, Bob Kahn, Severo Ornstein, Ben Barker, Frank
Heart, and to some extent Hawley Rising.  I can't exactly remember when
Alex McKenzie got involved but relatively early, nor can I remember
exactly when Ben Barker began packing the project with his Harvard
radio station friend, e.g., Joel Levin, Mike Kraley, and Marty Thrope.

During this period before I left for Norway in September 1970, the
amount of non-work activity in the BBN office increased dramatically
when BBN closed its Los Angeles computer systems activity and Truett
Thatch and Smokey Wallace relocated to Cambridge.  Smokey spent the
year or so he was at BBN in Cambridge before returning to California
primarily learning to play the guitar (with the occasional belt of red
wine from the bottle he kept in his office).

[Acknowledgement:  Bernie Cosell helped me remember some of these
details.]