Lawrence Moser Breed 1940-2021, A Personal Recollection

[Bob Smith]

Larry died peacefully with his family nearby on Sunday, May 16, 2021.

I met Larry in 1971 upon joining STSC where he was Vice President for
Systems. We hit it off right away as we shared a love for many things
including word games and puzzles, and of course APL.

Before I joined STSC, Larry had achieved a significant milestone in 1972
by jointly writing one of the world's first worldwide email systems,
named Mailbox, or 666 BOX to its users where he was known as LMB.

The annual Dictionary Rally was started by three of my friends to which
Larry caught on immediately, eventually winning the competition along
with his first wife Donna, an unheard of twice, undoubtedly because of
his keen sense of detail and love of words. He started the meme of Lost
Positives where we used his 13 volume Oxford English Dictionary to look
up “gruntled” where we found that “dis” is an intensifier, not a negater.

He shared with me many stories of the early days of APL design
discussions. Once, the group was deciding whether to keep the symbols
for the “and” and “or” functions since their result was duplicated by
“min” and “max”. Larry settled the argument by noting that those two
sets of functions had different identity elements, and so separate
symbols for those Boolean functions remained, which, later on, pleased
me greatly.

I remember that when he learned that he shared the 1973 Grace Murray
Hopper award for his work implementing APL\360, we bundled into my car
and brought back bottles of cold champagne so we could celebrate
properly in the middle of the day.

After he went back to IBM, we kept in touch, often exchanging small
gifts, such as a beautiful nautilus which still sits on my book shelf,
along with a very early plot of a prime spiral.

At some time, I made it a point to visit Larry and Beverly annually,
just to keep in touch. In time, his health deteriorated, but I was able
to convince him to attend the Minnowbrook APL Implementors Workshop in
2017. He flew to DC and we drove to the workshop in upstate NY where he
showed off more prime spiral plots as well as a collection of Dictionary
Rally dictionaries and instructions over the years. The 450 miles of the
trip up and back with Larry melted away as we reminisced about the many
good times we shared.

Due to ill health, he, Jim Brown, and Roger Hui couldn’t make it to the
2019 Minnowbrook gathering, so we asked the attendees to say a few words
of encouragement to each of them which Jon McGrew adeptly videotaped and
sent off.

Throughout the time I knew Larry, he was my mentor. I looked up to him
so much that I followed him into becoming an APL implementor and
language designer, all of which has given me such great pleasure over
many years. I can fully appreciate how he so much enjoyed those roles.

To say I miss him greatly is an enormous understatement.

Please share your own thoughts of your experiences working

--
_________________________________________
Bob Smith -- bsm...@sudleydeplacespam.com
http://www.sudleyplace.com - http://www.nars2000.org

To reply to me directly, delete "despam".

[Kerry Liles]

I met Larry when I attended an APL planning session in 1972(?) that was
held at what I believe was his parent's cabin high up in the Sierra
Nevada mountains in California... Bob Bernecky and I traveled there
together from IPSA, I recall Jim Brown being there and a number of
others... Larry's wife cooked all the meals with the provisio that
someone else from the group dealt with cleaning the dishes :) We all
took turns washing up.

Best memory: Larry's daughter (Emily??) was about 4 at that time and I
remember her taping Larry on the shoulder [when he was dialed into STSC
with an acoustic coupler modem using a spinwriter AJ terminal I think] -
she asked if she could get on next to check her email (!)

Wonderful family and Larry was a great inspiration. I felt that every
time he spoke it was clear he had forgotten more about computing than I
was likely to every know.

Rest in peace; condolences to his family.

[Rav]

On 5/19/2021 9:40 PM, Bob Smith wrote:

I met Larry in 1972 at STSC in Bethesda, MD after graduating high
school. He was one of a number of brilliant people there who inspired
me and helped me broaden myself both professionally and personally.

I remember there was a storage room containing office supplies and STSC
informational material, including bios on the senior staff there. One
was of Larry, containing a picture of a totally beardless and
short-haired Larry, looking (as I remember it) very, very young. I wish
I had saved that picture.

The following year when I was at college I wrote Larry a letter, which I
don’t have a copy of. Among other things, apparently I made a number of
suggestions about compiling APL, which although I wouldn’t have had the
first idea of how to implement, I would certainly have appreciated.
Larry mailed me back a hand-typed letter, which I do still have the
original of. I think his response is interesting.

It begins “This is a genuine offline letter -- the system’s been down
for several hours because of power failure, and I'm using the powerful,
flexible IBM 2741 in Local Concept Mode (or whatever IBM is calling it).
Note the low number of typos -- my specialty is ability to switch
finger-reflexes from APL to Courier to keypunch on short notice.” “Your
four ideas make sense (although deleting comment lines would probably
mess up programs written by people like [name deleted] who believe in
absolute line numbers) and could probably be incorporated in APL/360. I
suspect you haven't considered some of the problems with 'compilation',
though, whatever that means. (Back in the good old days I used to know,
but with APL I'm not so sure.) It sounds so straightforward to say ‘type
and rank must be known at compile time‘ -- but just what do we do with,
say, the argument and result, respectively, of FAPPEND and FREAD? Those
functions must be able to handle any APL quantity! Phil Abrams and I are
starting to consider seriously the problems of (ahem) efficient APL
execution; so far, it looks like whatever clever tricks we come up with
we'll need a full APL interpreter at run time, to fall back on when
compiled code fails us.”

Commenting on my major and extra-curricular activities, he approved and
said that they would “probably involve you in a far greater breadth of
activities than would computer-bumming. I am not enthusiastic about
computer science undergraduate majors; it just seems there must be more
significant things to be learning at that age. I know computers can be
lots of fun, but I also know that they're an unsurpassable temptation
for flunking German!”

Also apparently in my letter I reported that my first Fortran program
had compiled without errors and I was finding it all very interesting
(yes, I learned APL first, in high school, THEN learned Fortran and
assembly language. Boy did THAT make me appreciate APL even more.). I
love his response, particularly considering APLs longevity (today):
“Your comment on Fortran's acceptability is depressing … it merits
Twain's comparison to a dog walking on its hind legs. I earnestly hope
that your praise is motivated by, say, the understandable exhilaration
of getting your first program to actually run. No, this isn't my APL
chauvinism coming out -- it's more a belief that in a field developing
as rapidly as computer languages, a seventeen-year-old language just
can't be good. We've learned too much in the meantime.”

Goodbye, Larry. And thank you. Paul

[Quadibloc]

On Monday, May 24, 2021 at 8:44:30 AM UTC-6, Rav quoted, in part:

> it's more a belief that in a field developing
> as rapidly as computer languages, a seventeen-year-old language just
> can't be good. We've learned too much in the meantime.”

It is true that there are some good things about Pascal or Algol that were
missing from Fortran IV/Fortran-60.

Fortran-77 stole them, though.

There are certainly many languages that are more ambitious than Fortran
or BASIC. Ada, Modula-2, Algol 68 and APL all come to mind. But they haven't
taken the world by storm; in various ways, they were all _too_ ambitious, and
thus they weren't serviceable tools for those who just wanted to get work done.

Of course, nowadays we have languages like Python - which _is_ used quite a
lot, and which does get rave reviews.

It's fast, and slated to become faster - and it has libraries which make it applicable
to a wide variety of purposes.

So maybe we are now making actual progress in the field of computer languages.

John Savard

[J. Clarke]

On Tue, 25 May 2021 00:15:29 -0700 (PDT), Quadibloc
<jsa...@ecn.ab.ca> wrote:

>On Monday, May 24, 2021 at 8:44:30 AM UTC-6, Rav quoted, in part:
>> it's more a belief that in a field developing
>> as rapidly as computer languages, a seventeen-year-old language just
>> can't be good. We've learned too much in the meantime.”
>
>It is true that there are some good things about Pascal or Algol that were
>missing from Fortran IV/Fortran-60.
>
>Fortran-77 stole them, though.
>
>There are certainly many languages that are more ambitious than Fortran
>or BASIC. Ada, Modula-2, Algol 68 and APL all come to mind. But they haven't
>taken the world by storm; in various ways, they were all _too_ ambitious, and
>thus they weren't serviceable tools for those who just wanted to get work done.

Ada, Modula, and Algol 68 didn't make it easy to write programs. APL's
big shortcoming has always been price and accessibility. APL was
always a "serviceable tool for those who just wanted to get work

done".

>Of course, nowadays we have languages like Python - which _is_ used quite a
>lot, and which does get rave reviews.

And with Numpy is actually a viable substitute for APL.

[Lee C.]

Curtis Jones posted a note about Larry's passing on the Bay Area APL list. I am a lapsed APL fanatic having written my first APL program in the late-1960s in the 7th grade, and hacking a lot of APL in high school and college in the 1970s.

But, Larry and my paths did not cross until the 20'teens after I "retired" from Intel (meaning "You're too old now, take this money and leave.") and was working my retirement job at the hardware store in downtown Palo Alto. We became acquaintances when Larry would come in to purchase supplies for the Burning Man project he was working on. We knew each other well before I realized he was *THE* Larry Breed of APL lore. We never really talked in depth about APL, which I regret. But, always had great conversations about his fascinating BM and other projects. It gave me great satisfaction hanging out at the service desk in the back of the store or fasteners department, discussing solutions for the current obstacle he was working thru on the latest project. I'm a better person for having known and learned from Larry. And a better community for Larry's presence. God speed friend and RIP.

Lee Courtney

[Charles Brenner]

Thanks Bob, and others, for your stories. Add Larry Tesler and me to those for whom Larry was number one mentor.

I think my first encounter with Larry Breed was when he asked me to wind his watch. It was in the make-ready room to Stanford's Burroughs 220 computer in Encina Hall. It was about when I was finishing high-school, spring or summer of 1961. I had access to the Stanford computer through Prof. Forsythe who had delivered a remarkable lecture on how they work to our high-school science seminar.

Larry's right hand was in a cast. His watch was on his left but being left-handed he always wore it that way, not because of the cast. The cast was of course the reason he couldn't wind his own watch. Still, if you think about it it's an odd request in the circumstances. Anyway, winding a watch isn't difficult so I did, and moreover, noticing that the watch's time was off by about 6 hours I added the extra favor of correcting it. Larry objected and made it clear that he was not grateful. By way of explanation he pointed out what I'd noticed but not thought about. His minute hand wasn't synchronized with the hour hand in the usual way but rather Larry had re-positioned by just enough that the two hands would coincide not at the usual 12 o'clock but rather at 6:30. That allowed him to wear the watch upside down (why?) on his left wrist, so of course he ignored the numbers on the face and preferred to have both hands pointing up at 12 sharp.

I was duly chagrined.

However, forty years later, no longer as gullible as at 16, I asked Larry if he'd trapped me intentionally. He agreed. So not for the only time a Breed incident had its denouement decades after.

C