-
- Lynn Conway's
Retrospective
- Lynn Conway [Update of 2-25-09.]
- Copyright © 1999-2003,
Lynn Conway.
- All Rights
Reserved.
-
-
-
-
- PART III: STARTING
OVER
-
Starting over - - "going stealth" - - - settling down
in my new life - - - making new friends - - - dating men and
having fun - - the effect of the gender transformation is almost
like a miracle to me - - - I can finally just be myself, in a
body that feels great to be in, and openly love and be loved
simply for who I am - -
-
- Some nice career successes occur at Memorex - - I begin working
as a computer architect again - - - ah, but then project cancellation
- - - time to make a big career decision - - - great fortune
at joining PARC - - then amazingly - - - being mentored by Bert
Sutherland - - - meeting Ivan Sutherland, Carver Mead - - - getting
involved in some of the most exciting research going on in computing
at the time - - seeing an opportunity for creating new design
methods - - - innovating the VLSI system design methods - - -
creating a paradigm-shifting VLSI systems textbook - - - getting
bolder and more confident in my career, and taking many daring
steps along the way - - - and by ten years after my transformation,
I'm suddenly on the threshold of career fame - - whether I'm
ready for that, is another question - - - because I must live
in "stealth mode", and have to worry about losing my
career opportunities if I were ever publicly outed - -
-
-
- You miss 100 percent of the shots you never take
-
- - Wayne Gretzky
-
-
-
-
-
-
-
-
-
-
-
-
-
- The best way to predict the future is to invent
it -
- Alan Kay
-
-
-
- 5. STARTING ALL OVER AGAIN:
-
- - - - The late 60's were a remarkable time of new opportunities
for women - - - the 50's seemed like a distant past - - - many
companies were trying to bring many more women into regular professional
positions that had previously been predominantly male-only -
- - the period of quotas and affirmative action hadn't arrived
yet, so there wasn't much backlash from men - - - after all the
numbers of women in most places was still small - - - however,
the opportunities were there at just the right time for me -
- - especially in the then wide-open field of computer programming
- - - in fact, many of the new computer programmers were women
who learned their skills "on the job", much as the
early webdesigners did in the 90's - - -
-
- - - - it was in this dramatically changed context for women
at work that I got my first job as a woman at CAI as a programmer
- - - a number of other young women had also recently been hired
by CAI, and we all became good friends - - - this was great,
because it gave me a chance to immediately build a circle of
regular gals as friends and buddies - - - for socializing, shopping,
and having fun together - - - and this helped me stabilize my
life a lot and quickly reduced my feelings of loneliness - -
-
-
- - - - I worked on one timesharing project at CAI, learning
about time-sharing operating systems on the job, and making some
useful contributions by figuring out why their stuff wasn't working
right - - - once I was making some money again - - - I had moved
from San Francisco to a tiny one-room apartment down on the Peninsula
in Sunnyvale, and re-entered "the real world" as Lynn
Conway - - -
-
- - - - CAI was a good stepping stone - - - I was soon recruited
with several other women programmers from CAI to go work at Greyhound
Time-Sharing (GTS), which was just starting up - - one of my
programmer friends at work named Cindy and I decided to get an
apartment together as roomates - - - Cindy was a very beautiful,
friendly, fun-loving gal and we hit it off great together - -
- we found a really nice two-bedroom apartment on Middlefield
Road in Mountain View, where we lived together for a couple of
years - - - it was a new complex with nice grounds and a swimming
pool - - - it felt really great to be in decent housing again
- - -
-
- GTS was folded later that summer by the parent company, but
they were good about it, and we were given some time to look
for new jobs this was a huge scare for me - - - I thought good
grief, if I can't find a job again, I'm going to lose out after
all - - - however, the timing turned out to be quite expedient
- - - my headhunter put me onto to a really good job at Memorex
Corporation, which I landed in September 1969 - - - finally the
darkest days of my early life were now over, and the sun was
about to shine a little! - - -
-
-
- Memorex Corporation
-
- - - - In 1969, memorex was an exciting start-up that had
"gone critical" and become a substantial force in the
IBM compatible disc-drive business - - they would soon even enter
the computer business (small business systems) - - there seemed
to be possibilities for an exciting future there - - - although
I didn't go into a lot of detail about my work at IBM, my background
in system architecture, digital design and simulation work gave
me a lot of credibility with the hiring managers and personnel
people - - and I got a good offer from them - -
-
- - - - and the offer was still there after I informed the
personnel department of my medical history, and showed them full
documentation regarding what had been done, including a copy
of Dr.
Julian Pichel's Psychiatric Report - - - that independant
psychiatric examination by someone not involved in my case management,
and which was supportive without hesistation, was very useful
in making the case - - besides, no one at Memorex expressed anything
remotely like the concerns that the IBM corporate folks had obsessed
about just the year before - - -
-
- - - - these were younger folks in a recent start-up company
- - - folks who had met me personally, and who simply listened
carefully to my detailed explanation of what had happened to
me - - - they were used to making up their own minds about things,
rather than relying on prevailing prejudgments - - - so I joined
Memorex - - - and was soon to begin a really good job, in a really
good company - - -
-
-
- Restarting my engineering
career and my social life
-
- - - - but basically, I had to start all over again - - -
for one thing, the IBM-ACS project had been so secret that the
project itself didn't have an identity among engineers in the
valley, much less any work that I'd done there - - - and there
wasn't much I could say about what I'd done at IBM anyways, for
fear of revealing my gender transition - - - then too, I needed
to develop a wide range of new hands-on technical skills in order
to work within the fast-paced, start-up type of environment -
- - so I had to start all over pretty much from scratch technically,
and prove myself all over again, first as a systems programmer,
then moving on to work as a digital designer and computer engineer,
and then finally getting a chance to work on computer architecture
again - - -
-
- - - - this wasn't easy and was sometimes pretty scary - -
even though I'd had some realistic social and medical/physical
experiences when I was 20-22 years old, I was now experiencing
a complete and profound new internal and external reality, going
through what amounts to a "second puberty" - - by now
the overall transformation was bringing on a deluge of new experiences
at work, and especially in my personal life and in my own internal
physical perceptions and emotional experiences - -
-
- - - then, in 1970-71, the case of April
Ashley broke into the news - - April was a beautiful English
socialite who had been a famous model in the early 60's and who
had then married a prominent peer - - however, she had been born
a boy from a poor family, and had worked as a female impersonator
at Le Carrousel in Paris - - - she had become one of Georges
Burou's earliest patients (I understand she was SRS number 6
by Dr Burou in Casablanca on May 12, 1960), and thus was among
the first tiny band of transsexual women to undergo modern "sex-change"
surgery - - - see
recent news article about April - -
-
- - - living in stealth mode, April was "outed" in
1970, and her husband's wealthy family forced the case into the
divorce courts where the marriage was declared invalid - - the
courts even declared that April was "still a man" since
she was XY genetically. This case set a disastrous precedent
that has affected post-op women in the UK ever since (they are
considered by the legal system there to be males, and cannot
marry "another man") - - and it terribly frightened
me - - - the idea of being "outed" and somehow declared
to "be a man" was an unthinkable thing to be avoided
at all costs - - so for the following 30 years I almost never
talked about my past to anyone other than close friends and a
few lovers - - -
-
- - - - by '70 I'd only just begun to make some new friends
- - - and although I had a few friends to socialize with and
gain emotional support from, I still felt withdrawels and loneliness
and alienation from the huge transition I'd just gone through
- - - I also missed seeing my little girls a lot, but tried not
to think about this if at all possible - - - I also found myself
slipping in and out of bouts with shyness and social withdrawal
because of my terrible fears of being outed - - - however, I
worked very, very hard at getting things going, while trying
not to make too many mistakes - -
-
- - - - having Cindy for a roomate was wonderful, because we
could go shopping, go out socializing, and do a lot of fun things
together - - - she helped me grow out of my shyness more and
more - - - and within a couple of years after SRS, I was settling
into my new life pretty comfortably - -
-
- - - - my career also took a great turn for the better when
in '71, after a successing of increasingly more responsible positions
at Memorex, I was selected as processor architect for a new small
business computer system - - the Memorex 30 - - -
-
-
- The Memorex 7100 Processor
-
- - - - the Memorex 7100 processor for Memorex 30 System, a
new low-end model for Memorex's line of computers - - - aimed
at competing with IBM's System 3 - - -
-
- - - - Milt Gregory pulled together and led the matrixed project
team - - - the fast development cycle - - my initial architectural
ideas - - this was to be a lean, mean, low-cost, no-frills machine
- - - the use of a simulator to cohere the design - - our effort
to get on top of all the newest parts due out from various suppliers
- - - in 9 months we went from blank piece of paper to a working
manufacturing prototype - - a microprogrammed processor in TTL
technology - -
-
- The Memorex 30 Team:
-
- Milt Gregory, Anthony J "Tony" Miller, Lynn Conway, Roger Stallman, Gary
Yee, Bob Peterson, Al Hemel, Dick Hoenle, Bob Holland, Don Pesavento,
Glenn Ewart, Dick Chueh and Barbara Baird.
-
-
- - - - this project developed into a wonderful group experience
- - with a great bunch of people who were fun to work with, and
who were really sweet to me - - - I learned a lot on this project
about hands-on engineering of real products - - what the digital
designer's life in the trenches was really like - - - and I really
enjoyed the exciting team-effort aspect of the project - - -
Milt was a great leader, and I learned a lot about team leadership
by watching how he did things - - - also, late in the effort,
the team worked with Barb Baird to produce a manual documenting
the design - - fortunately, I still have a copy - - -
Update of 2-25-09: I've just scanned and posted the Memorex 7100 System
Reference Manual online. You'll find PDF's of the complete manual and separate
chapters at the following links:
Memorex 7100 System Reference Manual,
Front-matter,
Introduction,
System Architecture,
CPU Architecture and
Microprogramming, CPU
Logic Design, Memory
System,
System
Control and Display Panel,
I/O System,
Physical Description,
Power Supply,
OPSYS1 Emulation Package,
MRX30
Emulation and Performance,
MRX30 Phase
"0" Cost Estimate.
-
-
-
-
- - - at this point, things were going great - - I was making
excellent money, and was accepted as a senior technical peer
by leading engineers in the company - - - I'd also gained confidence
as a technical project leader, while guiding the processor development
portion of the overall project - - - and in 1971 I was promoted
to Senior Staff Engineer - -
-
-
- Having Fun
-
- - - - I bought a nice little condo in Los Gatos in '72 -
- - conveniently located at 420
Alberto Way (#25), near to downtown - - - by now my past
troubles had sort of melted away, and my inner perceptual, emotional
and physical setting had settled down and stabilized into its
wonderful new form - - - I'd also made significant gains in social
adventurousness, and was freely enjoying an active social life
- - - I traded in my old Volkswagen Beetle on a new '72 Datsun
240Z - - - it was red with a white interior; what a car! - -
- that thing helped propel me further into some wonderful new
experiences - - -
-
-
- Lynn's Datsun 240Z in front of
her condo building, in 1972.
-
-
- - - - I loved dancing to rock & roll, rhythm & blues,
and country music at the roadhouses like "Cowtown"
down in San Jose - - - and barhopping in general - - - I was
now dating men regularly - - I got started by frequenting the
various singles bars and by using the then-new "computer
dating services" - - - and from time to time I began having
really wild fun - - the surgery had finally released me to fully
enjoy lovemaking - - I could wrap my legs around a man and make
love with wild abandon - - without elaborate preparations, without
fear, shame or embarrassment, and without hiding anything - -
- and my body was now beautifully soft and rubbery, and sweetly
sensitive and responsive to a man's touch, due to the female
hormones - - - my god, what joy this brought me - - -
-
- - - - I also now had two
beautiful Siamese cats , Samantha and Rapunzel - - I've always
loved cats, and these two sweet little critters were to be my
faithful companions for many, many years into the future (Sams
lived to be 21 and Punzel lived to age 23!) - - -
-
- From this point on, my personal life was much like that of
any other single woman in her mid-thirties, full of the same
challenges and rewards that lively women of that age experience.
For many years I dated regularly, and usually had a couple of
somewhat steady boyfriends with whom to share interests and have
fun - - - here are some photos from this period:
-
|
- Lynn in her
- '72 Datsun 240-Z
-
- 1973
|
-
-
 |
Lynn at a car show
1973 |
-
-
|
Lynn's roomate Cindy
1971 |
-
-
-
- The 4004 Appears
-
- - - in 1971 something very significant happened in the valley
- - - the Intel 4004 appeared and soon began to get into all
sorts of applications - - a blockbuster event for digital system
designers - - I talked to anybody who knew anything about it
- - took short intensive courses at Santa Clara University -
- tried to find out what was going on - - the processor was very
simple architecturally - - - but the design details and process
were very unclear - - do architects have to know anything about
LSI to do this stuff? - - apparently not - - it's a big team
effort involving architects, logic designers, circuit designers
and layout designers - - - it's complex and costly to design,
even for such a simple processor? - - - but makes sense if there's
lots of applications because it's so inexpensive to replicate
- - -
-
-
|
Photomicrograph of the Intel
4004 chip - the first microprocessor |
|
-
-
- Project Cancellation
-
- - - - but then there are rumblings - - business problems
- - - gads, the MRX 30 project is canceled! - - - not only that,
but Memorex announces that it's getting out of the computer business
- - - this was a terrible disappointment - - I felt like I had
the flu for the next month - - - now what to do? - - - we went
into hold mode for a while, and then there were layoffs - - I
survived those, but finally couldn't see a future at Memorex
- -
-
- What Next?
-
- - - What to do? - - - I had to get a grip on things and decide
what to do next - - - I finally signed up a headhunter and went
looking for a microprocessor to architect! - - heck, why not?
- - there were very few folks in the valley at the time who knew
how to do computer architecture - - - so I'd guessed that those
architects already doing the microprocessors didn't know much
about the underlying technology either - - -
-
- - - - this proved to be a good guess - - - I interviewed
with and got a great offer from Fairchild - - - but then my headhunter
mentioned that Xerox had started a new research outfit in Palo
Alto that might be interesting - - I went to check it out - -
- I could hardly believe what I saw on that interview trip -
- PARC in 1973 was already an astounding place - - the mission
- - the people - - the computing environment - - the special
feeling of a different, exotic technical culture - - - I got
an excellent offer from Xerox - - - and the offer was still on
the table after I'd fully informed the personnel manager about
my past history, using all my medical documentation plus Dr.
Pichel's Report - - -
-
- Note: By 1973,
transsexualism was much more widely known about than in '68,
especially in the Bay Area. In the five years since I'd been
fired by IBM, Dr. Benjamin's book had had a great impact in the
medical community. Stanford University Medical Center in particular
had started an exploratory program in the late 60's to evaluate
transsexual cases and to develop standard protocols for treatment.
The program resulted from knowledge of successful cases, including
mine, that Dr. Benjamin had diagnosed and then guided on through
SRS and transition in the 60's.
-
- Back in 1968, I'd been among the first few
hundred U. S. citizens to undergo SRS, and almost all of them
had gone abroad for their surgery. The majority were entertainers,
sex-trade workers, and others who could pass and find marginal
work as women back then. That's why there hadn't been any precedents
for IBM's corporate response to my transition. Such a thing just
hadn't happened before in IBM, or in any other major corporation
that they knew about, and there was no script for handling the
situation.
-
- However, during the late 60's as Dr. Benjamin's
work became known, many more transsexuals sought medical help.
According to Dr. Benjamin, by 1973 almost 2500 Americans had
undergone the change - - - and with each passing year a higher
percentage were educated and professional women who were able
to maintain some degree of career success afterwards - - - by
then the Stanford efforts were generally known about by the public
in the Bay Area, having received occasional news attention -
- this started to open folks minds to the idea that there was
a rare gender-transition disorder that an elite organization
such as Stanford took seriously, and wanted to study and help
- - - it's possible that this Stanford involvement in treatment
of transsexuals helped sway the Xerox personnel department in
my favor during the hiring process there - - - I kept in touch
with Dr. Benjamin for many years after my transition, usually
seeing him for a lunch or dinner at least once each summer when
he was in San Francisco. It was wonderful to be able to share
in his happiness in knowing that his work was having a lot of
impact and helping many desperate people, and that he would eventually
be recognized as an great medical pioneer.
-
- - - - I did get another big scare during
the summer of '76 when Renee
Richards, a then recently post-op transsexual, was "outed"
by national media (her story is documented in her '83 book, Second Serve).
A huge controversy raged about her playing in tennis competitions
as a woman, and people made fun of her as they had of Christine
Jorgensen years before - - - even more so in Richard's case because
she was very strange-looking (unfortunately for her she had very
masculine facial features) - - - I feared the highly publicized
controversy would cause me to be scrutinized more closely by
colleagues, and that I might possibly be outed then too - - that
didn't happen, but my old fears were surfaced again and held
me back from being more "out front" and in the public
eye as my VLSI work became successful several years later - -
-
-
- - - - medical progress continued, and by
1979 Stanford had developed the protocols that have been used
to this day for TS case management and gatekeeping to SRS - -
By the late '80's, the numbers of cases being processed in the
U.S. rose up to its current, ongoing level of about 1500 cases
per year going through MtF SRS. Perhaps another 500 or so go
outside the U.S. for their surgery. That's a tiny number when
spread around such a large country; however, it's still a lot
of people who must undergo a frightening, painful and often desperate
struggle to have any chance at a real life.
-
- By the year 2000, there were approximately
60,000 people in the U.S. who had undergone SRS (MtF or FtM).
The vast majority were in stealth in their new lives, and simply
blend in unnoticed among the rest of society.
-
-
- - - - now I had a very important decision to make - - - should
I take the offer from Fairchild to do their next microprocessor
architecture? Or join Xerox in Palo Alto to do research on special
purpose architectures for image processing? - - - so many factors
were in play in this decision - - - both seemed like outstanding
opportunities - - - but the desire to work in an exciting, creative
research environment again, as I had at IBM-ACS, finally drew
me to Xerox - -
-
- - - - there had been big turning points in my life before
- - and here was another one - - I faced a decision point - -
- going from a place of comfortable career success and an increasingly
active, happy social life - - to dropping into the intellectual
supercooker that was Xerox PARC - - a place of unlimited possibilities
for creative expression - - - but at a price of all-consuming
work for quite a few years - -
-
- - - - it's hard to say whether the eventual price was too
high, or not - - shouldn't I have stepped back, and considered
more closely all the wonderful new alternatives that were now
fully open to me in life? - - - why did I only consider the two
major career opportunities? - - -
-
- - - - was it fear of financial insecurity? - - - this had
been a big source of anxiety all during my transition, because
of the huge unreimbursed medical expenses and also the expenses
for supporting Kelly and Tracy - - -
-
- - - - or was it a lack of confidence in the future for my
personal life? - - - although I was feeling really good about
myself and enjoying life - - - I had a growing realization that
while some men liked to date and enjoy intimate fun with transsexual
women, who are often very passionate partners, few men would
knowingly take a TS woman for a wife due to social stigmatization
- - -
-
- - - - I'd begun to worry that although I could have a fun
social life and sex life, I might not be able find a mate to
share life with - - - I'd learned that once you were emotionally
involved with someone, in fairness you had to tell them about
the past - - and in most cases you'd then either lose them or
the relationship would shift into a different and less-serious
phase - - after having this happens to you several times, you
can get a bit burned out and tend to give up on love and just
settle for less meaningful affairs with men who "don't know"
- - - this has always been a big quandary for TS women (although
things gradually became a lot better by the 90's as the condition
became less stigmatized) - - -
-
- - - - or was it a kind of overcompensation, the need to make
enough contributions to society to counterbalance in the minds
of others their potential reactions to my past? - - - such overcompensation
is common among post-op TS's - - - a lingering result of very
low self-esteem, in spite of any later successes, going all the
way back to the humiliation and derision I faced in my early
years right on thru to my firing by IBM - - -
-
- - - - or was making creative contributions a substitute for
being unable to bear children (this is a sensitive topic that
I don't like to dwell on - - to this day it can bring tears -
- especially because of the forced loss of all contact with little
Kelly and Tracy, who had seemed like my own babies to me at the
time of my transition) - - -
-
- - - - or was it that, in addition to everything else, I had
a wild passion for the life of the creative mind too? - - - who
knows - - but the decision was made, and on I went to the Xerox Palo Alto
Research Center - - - and a period of intense focus on my
research career - - -
-
- [return to Contents]
-
-
- 6. XEROX
PALO ALTO RESEARCH CENTER (PARC)
-
- The SIERRA Project
-
- At Xerox, I initially reported to Mike Wilmer, in PARC's
Systems Sciences Laboratory (SSL), to work on a special purpose
system architecture for image processing - - the target application
was to try to create a compound OCR/FAX system - - - the "Sierra
project" - - - I innovated a cool, though on reflection
somewhat bizarre, architecture for this system - - -
-
- - - - after considerable toying around with various image
processing variants - - - Jorge Hernandez and I design and build
a particular hardware prototype - - I conduct extensive simulation
explorations of the OCR methods in parallel with hardware design
- - we get the thing running - - it really works! - - and in
later retrospect, I gained many important meta-architectural
insights from this experience - -
-
- - - - but we began to take flak from folks in the Computer
Science lab (CSL), who sniped at this project for being too complex
and costly - - - I don't think the presentations I gave about
Sierra in "Dealer" were very effective either, and
that didn't help - - I was still shy enough to get a bit sweaty,
and verbally hesitant, when talking in front of a group - - I
was in awe of what the CSL team had accomplished with the Alto
and the new form of distributed computing environment at PARC
- - - under the visionary leadership of Bob
Taylor - - - I hoped to get their respect - - but it wasn't
to be based on Sierra - - - it was gradually dawning on me too
that even with major re-engineering, any Sierra product prototypes
would be way too costly at the time - - -
-
- - - - about this time, in late '74 or early '75, we got a
new lab manager in SSL - - - his name was Bert
Sutherland - - - I recall a critical meeting with Bert, where
I presented the Sierra architecture to him and to his consultant,
the famous computer architect Wes
Clark - - - my impression was that Bert and Wes thought I'd
done some interesting (and certainly different!) architectural
work - - - hmm, would this mean that Sierra could continue? -
- -
-
- - - - I was pretty happy about how things were going for
me outside of work at the time - - - I'd was doing lots of fun
things, including making backpack trips to the Sierra Nevada
again - - -
-
- - - - I found that it was easy to meet some nice guys by
getting involved again in hiking and backpacking - - - it was
a great way to go on long dates and have something special to
share with your partner - - - and since there were more guys
than gals involved in those sports back then, it helped improve
the odds of meeting someone nice - - -
-
- - - - I was hoping to continue combining a fun personal life
with reasonable career success at Xerox - - - but the Sierra
project at Xerox was just too great a technical reach at the
time to be commercially feasible, given the low density and high
cost of the many components it would take to build it - - - and
it wasn't long before Bert canceled the project! - -
-
-
- Lynn in 1975, on a backpacking
trip to the Sierra Nevada with a boyfriend,
- gets ready for a day of hiking
and peak climbing (shown here near Ragged Peak).
|
 |
 |
-
-
- Lynn at a tent site on the approach
to Mount Ritter and Banner Peak,
- while on another backpack trip
with a boyfriend, in 1975.
-
-
- SIERRA Aftermath
-
- - - - this cancellation of Sierra was very unsettling - -
as a system architect and designer this was the third major project
canceled out from under me - - also, I wasn't sure how to save
my image within the competitive PARC environment - - - given
this "failure", I figured I was probably doomed to
be an underdog there - - -
- - - - about this time, I had moved in order to be closer
to Xerox PARC. I'd been driving my 240Z on mostly windy back
roads from Los Gatos to Palo Alto to get to work, and although
it was a fun drive, it was very time-consuming. With a now-secure
position, I was able to afford a nice little home at 1538
Winding Way in Belmont, CA. It was high on a hill overlooking
the bay - - - and the drive to PARC was shorter and was against
the main traffic and thus a lot easier - - - this would turn
out to be my cozy home during the period of intense work ahead
at PARC - - - although right now I worried that this would be
a repeat of my experience at Memorex, where I'd just bought a
condo and then changed jobs to a location a lot further away
- - -
-
- - - what a roller-coaster ride - - I was feeling pretty low
about my work right now - - although in retrospect this was,
and still is, a very common experience for high-tech workers
in Silicon Valley, so I shouldn't have taken it so hard - - many
folks spend their entire careers there, and never really "pick
a winner" - - never become part of a really successful project
that makes it to market and succeeds there - - -
-
- - - however, I couldn't have guessed what an incredible intellectual
adventure was just around the corner - - and what a great research
leader and wonderful personal mentor Bert Sutherland would turn
out to be - - -
-
- - - - I owe Bert Sutherland so much for everything I've accomplished
since then - - -
-
- - - Bert introduced me and my PARC colleague Doug Fairbairn
to his brother Ivan
Sutherland, and to Carver
Mead, at Caltech - - Ivan and Carver were investigating the
longer-term potential of MOS-LSI - - - Carver had a device physics
background, and had done extremely important theoretical work
on determining the likely fundamental gate size limits for MOSFETs
- - his theoretical lower limits were amazingly small - - Ivan
had estimated and articulated the possible opportunity space
that such limits opened up for computer system designers - -
but also, given his deep background in system design, computer
graphics and CAD, Ivan raised the key question of how would architects
cope with the complexity of such very large scale designs - -
-
- Formation of the VLSI System
Design Area at PARC
-
- Doug and I team up - - as Bert Sutherland then forms the
new "LSI Systems Area" (later to be called the VLSI
Design Area) in his lab to "explore LSI design methods and
tools" - - this was a big step for me, as I became responsible
as Area Manager - - gads, was I really ready for this? - - -
I begin recruiting - - and Doug and I begin lots of interactions
with Carver and his students - - -
Photo of Lynn in her office at PARC (1977)
-
- - - - Carver had important Intel connections, having been
a consultant there since the early 70's - - and based on his
Intel experiences, he began teaching courses at Caltech on the
then current Intel methods for MOS-LSI circuit and layout design,
including examples of ALU designs, etc., the sort of designs
actually being used in the new Intel microprocessors - -
-
- - - - Carver's courses were totally unique at the time -
- there were no other university course anywhere that taught
students how to implement this level of logical functions into
MOS-LSI layouts - - - Carver also had access to Intel's and other's
fab lines, where occasional "test chips" could be dropped
in for processing, so some student projects could occasionally
be fabricated - - - but even though his students did more of
the overall design themselves - - and thus were often more efficient
in prototype design productions than existing industry teams
- - - Carver could see that no knowledge then existed to deal
with the eventual complexity of designs as chip technology moved
towards his theorized ultimate limits of density - -
-
- - - - Doug and I learn about all this stuff from Carver -
- and do some preliminary LSI chip designs ourselves - - - Carver
was a real character, very charismatic, and very passionate about
the potential of MOS-LSI - - he wanted to get more system designers
interested in it, especially given the work he'd done with Ivan
that revealed the eventual potential of the technology - - he
just didn't understand why there wasn't more support for research
into the design aspects of LSI - - - so he became very excited
by our enthusiasm as we got more involved - - he was also wowed
by the amazing computing environment at PARC (the Alto's, ethernet,
laser printers, etc.) and the possibilities that this then-unique
environment presented - -
-
- - - it seems strange now that almost no computer architects
had paid any attention to Carver before - - perhaps the idea
of working at the layout level, or even having to know anything
about layouts, seemed beneath most architects - - this may have
been a combination of looking down on the slowness of MOS technologies
vs the bipolar ECL technologies, and a holdover effect from the
old pMOS metal gate days, when "only peons worked on layouts",
cutting rubies, etc. - - - "after all, the architect could
just hand-off an RTL-level design to be implemented into logic,
and the rest was all optimizeable at each level by the different
specialists, wasn't it?" - - -
-
- - - - how could these architects have known that by exploiting
Robert
Bower's invention of the self-aligned gate and the new ion-implantation
techniques, the emerging nMOS (and later CMOS) technology was
now a totally different beast, and full of architectural promise
- - especially since it got faster and faster as it scaled down
towards the limits that Carver had calculated - - -
-
- - - in contrast, Doug and I had no such past preconceptions
against learning about the details of MOS technology and circuit
layout - - we were just so amazed to be able to get at MOS-LSI,
and design things in it, that we jumped in and learned about
it as fast as we could - - from Carver, from his students, from
taking short courses in all sorts of related areas (such two
or three day "intensives" were a key method for knowledge
transfer in the Valley in those days) - - - ( one of these courses
was an intensive on charge transfer devices given by Carlo Sequin,
a CCD pioneer from Bell Labs who'd just joined the faculty at
U. C. Berkely - - Carlo turned out to be an important partner
in our later efforts ) - - - and we learned quickly by doing
more and more design ourselves - -
-
- - - - with Bert's encouragement, Doug and Jim Rowson, one
of Carver's students, began work on the design and programming
of a graphical LSI layout system on the Xerox Alto - - this creation
of new forms of software tools for the Alto became the early
important rationale within Xerox PARC for the LSI Systems Area's
research - - this long served as a "cover story" -
- and was, for example, sort of tolerated by CSL (we were finding
new uses for the Alto) - - even as our work later expanded out,
bootstrapping to impact the wider arena - -
-
- - - - having designed digital logic in everything from relays,
to vacuum tubes, to discrete transistors, to SSI, to MSI-TTL,
the move to MOS design was yet another familiar technology transition
exercise for me - - probably like learing a new language for
some folks - - and, being experienced at multiple levels of design
(physical, circuit, logic, architecture), I studied hard to understand
those levels in nMOS-LSI too - -
-
- - - curiously, there also seemed to be a lot going on in
the Intel designs that wasn't directly stratified in traditional
architectural, logic, circuit and layout levels of design - -
- this reminded me of ACS, where much of the actual knowledge
at the hybridized interfaces of architecture, logic-design and
circuit design were not formally codified in current "texts"
- - this further complicated things - - there were a lot of hybrid
structures reminiscent of the relay switching circuits I'd studied
in Caldwell's book on "Switching Circuits and Logical Design"
at Columbia - - - was there some way to taxonomize, classify,
exploit and simplify all this stuff? - -
-
- The idea of "designing
design methods": the Mead-Conway collaboration
- - - then, reflecting on my earlier ACS design process efforts,
I began to sense some very weird possibilities - - - given Carver's
theories about the ultimate scaling limits - - perhaps MOS-LSI
really could open up the room for system architects to finally
realize their "architectural dreams" - - not just the
usual workstation or minicomputer scale of machines - - - but
for grand things like I'd experienced in the superscalar and
image processing architectures - - and actually get them economically
implemented - - but this could only happen if more digital
system architects knew how to think about designing in that medium
- - and had a faster, less expensive way to do it than to
involve an entire team of specialists (i.e., logic designers,
circuit designers, layout designers, process designers) for each
chip design - -
-
- - - - again, recalling my logic design standardization effort
at ACS, and the "designing the ACS design process"
proposal - - where I had viewed a "design level" as
subject to redesign, even to rather significant paradigmatic
change - - I began to recall and reflect on the Steinmetz and
Shannon precedents - -
-
- - - - especially Charles Steinmetz,
who was one of my true long-time heroes - - the immigrant and
social underdog - - - who, in spite of a major physical handicap,
rose to technical prominence at the great General Electric Company
- - who single-handedly had generated, shaped and channeled knowledge
about AC phenomena, via his symbolic mathematical methods, so
that the average engineer could work with AC - - his methods,
taught at Union College, in Schenectady, NY were largely responsible
for a huge wave of progress in the electrical industry - - it
seemed as if we were now at the threshold of something similar,
a similar opportunity and challenge - -
-
- - - - reflecting on that past history - - - and on how Carver
had been trying to proselytize current expert architects into
becoming LSI practitioners "by example", i.e., by using
existing, evolving methods, and creating examples of new, rather
clean architectures quickly designed by small teams - - I just
didn't see how that form of proselytizing would work - - - it
would only impact in an ad-hoc manner over a long period of time
- -
-
- - - - not only that, but I think we all intuitively perceived
that many more possible system architectures could be created
and be cost-effective in MOS-LSI than previously thought - -
- thus it would take many, many more architects to explore and
do this all this new stuff - -
-
- - - - so we couldn't rely on existing"advanced computer
architects" - - but needed lots more "regular digital
system designers" for all kinds of new systems in VLSI -
- so it didn't make sense to try to just win over existing architects
to do LSI - - - there wouldn't be enough of them - - somehow
what was needed were methods to create lots of new system architects
- - from scratch - - so, again, the Steinmetz story comes to
mind - -
-
- - - - I suggested to Carver that we deliberately design new,
simplified MOS-LSI design methods, deliberately aimed at
not just the current expert digital system architect - - but
more directly at even the "budding, novice architect"
- - making it so easy to get started that more of them would
try it, and work from architecture all the way to the layout
level - - Carver later coined the term "the tall thin person"
for this "new type of VLSI system designer" - - - Doug
and Jim could help forge new computer-aided design tools to support
this new sort of designer - -
-
- - - my idea was to deliberately distill and recompose from
all available nMOS logic, circuit and layout design methods the
simplest hybridized methods possible that would let a digital
system architect (for example, a Conway) jump right in and do
a whole system design from top to bottom - - i.e., the sort of
thing that folks like Lynn Conway would really liked to have
learned back when she first heard about the 4004 - - -
-
- - - - and also to adjust various viewpoints about "just
what a chip design was" - - getting it into 2-D overview
needed for the architect, rather than always focusing on 3-D
cross-sections for circuit details, and non-topologically-relevant
logic-gate diagrams traditionally used for logic design - - when
shifting from bipolar to MOS technologies, we could bring back
echoes of useful design thought-styles from relay switching-circuits
and vacuum tube circuits - - simplifying ideas - -
-
- - - thus a shared vision began to develop in our team - -
not just Carver's original "theory about eventual ultimate
limits" and "the challenge of complexity" - -
but my vision and approach on how to "deal with complexity
through deliberate methodological simplifications" - - -
and my concepts of how to craft simplifications that had a good
impedance match with the knowledge base of the average digital
designer of that time - - - we would create a new "VLSI
design methodology" - - -
-
 |
- One of the great heroes of the
early days
- of electrical engineering:
-
- Charles
Steinmetz
-
- His development of
- a symbolic method of calculating
- alternating current (AC) phenomena,
- simplified an extremely complicated
field, understood by few,
- so that the average engineer
- could work with AC.
-
- This accomplishment was largely
- responsible for the rapid progress
- made in the commercial
- introduction of AC apparatus.
|
-
- - - - well, just what did this VLSI design methodology begin
to consist of? - - - my idea was to initially only teach potential
designers about a very small key subset of available LSI
logic, circuit and layout methods - - - a key subset that would
cover the design of any digital system or computing structure
- - - a key subset that was internally simple, elegant, and easy
to teach to the existing digital design community - - -
-
- - - - we would use only simple two-phase clocking - - - use
simple dynamic registers - - - teach basics at the circuit level
to enable transistor ratio calculations, delay calculations,
establish fan-out rules, driving of large loads, etc. - - exploit
pass-gate logic in various classic architectural building-blocks
- - - use PLA's for any messy logic and for control logic - -
- use "stick diagrams" for initial system cells to
layout conceptual designs - - establish system-level structures
of register to register data paths, often with pass-gate logic
between register arrays - - with control lines perpendicular
to the data paths - - current density limitations - - power calculations
- -
-
- The innovation of scalable design
rules, based on the length-unit "λ":
-
- - - thinking about stick diagrams - - I pondered the difficulty
and computational complexity of trying to instantiate layouts
using typical layout design rules - - this was especially a problem
when using personal workstation-based graphics-layout tools like
Icarus - - - then I hit on the idea of using simple, scalable
design rules - - these became known as the so-called "lambda
rules" - - - this innovation greatly simplified the conceptualization
of chip layouts - - - and also dramatically reduced the computational
complexity of the layout descriptions - - -
-
- - - - Carver really liked the scalable design rules idea
- - - he helped refine some line width and spacing ratios, based
on estimates of future nMOS process evolution - - to make the
rules a bit more optimal and hopefully more "acceptable"
- - - we ended up with a just two pages of very basic, easy to
understand design rules, whereas most processes had about 30
pages with all sorts of arcane stuff in them - -
-
|
Plate 2. Mead-Conway Scalable
nMOS Layout Design Rules |
 |
-
-
|
Plate 3. Mead-Conway Scalable
nMOS Layout Design Rules (cont.) |
 |
-
- - - - however, these "non-optimal" rules later
became very controversial, and caused considerable backlash from
establishment figures who didn't seem to recognize the issues
of computational complexity, teachability, and time-to-design
tradeoffs that we were concentrating on - - designs created under
these rules looked so "clean and simple" that they
at first looked like "toys" to establishment folks
who didn't understand what we were doing - - - interestingly,
the rules gave up much less in area than most critics initially
guessed - -
-
- - - but the new design rules also opened up an important
new opportunity - - we could get access to various fab lines
by varying only one basic parameter, the length unit lambda,
for a given layout design - - we all worked to tabulate and simplify
the other factors at the fab line interface - - - and Carver
went on a quest to interact with all his contacts to set up cleaner
interfaces to a number of different fab lines, so as to increase
the opportunities for finding fab for student projects - -
-
- Carver coins the term "silicon foundry" to
promote clean interfaces with maskmakers and fab lines:
-
- - - - Carver coined the term "silicon foundry" for a line
that can be accessed in the new "clean way", i.e.,
simple logistical interface, standard ways to prescribe mask
polarities and alignment marks, simple scaling of the lambda
design rules, etc.- - and importantly, one definitely need not
jointly design and spec a process for each particular chip design
- -
-
- - - - the term foundry later became very controversial -
- - it really bugged lots of people, who didn't like it's apparent
dethroning of fab as the end-all, be-all of chipmaking - - (note:
the term is now in everyday use in the semiconductor industry)
- - Carver had a way of bringing notice to himself and
to our work by confronting traditionalists with what seemed
to be rather outrageous claims, and with inventive, but slightly
offensive, new terminology! - - Carver was always memorable!
- - -
-
- - - - Mead's growing interest in VLSI computer architecture:
It's interesting, given his background in device physics and
circuit design, Carver became extremely interested in computer
architecture, and wanted to "do it" himself - - - whereas
I, with my background in computer architecture, became extremely
interested in meta-architectural issues of selecting and reconstructing
the lower-level and mid-level zones of VLSI design to make them
into a "kit of parts that was more accessible to architects"
- - - we both seemed to want to learn and master what we already
didn't know much about - - to each of us the other area "seemed
cooler", at the time - -
-
- - - - therefore, in parallel with my obsessing on the design
methods at PARC, Carver focused intense energy on the architecture
and design of LSI microprocessors, the so-called "OM machines",
at Caltech - - - collaborating with his student Dave Johannsen
- - on a new machine, the OM-2 - - - OM-2 was just starting up
just as we finalized details of the design methodology - - -
Carver and Dave and several other students carried out the design
- -
-
- - - - fortunately, Carver was able to impose the new design
methodology being worked out at PARC on the OM-2 - - as a result,
the OM2 and all its detailed subsystems later became usable as
the classic microprocessor design example for Mead-Conway text
- - - somehow when conveying the methods to the OM2 team, Carver
didn't mention they were my idea - - - and when I visited Caltech
later, OM2 team members excitedly told me how they were using
"Carver design rules" as though I'd never heard about
them, much less invented them - - -
-
- - - - this caused a big flap between Carver and me, but didn't
resolve the issue - - - and "who was I" anyways - -
and we entered a phase where I got innovative ideas
and "worked them up in the back room" while he played
famous professor out on the public stage flaming about our results
- - - it was actually OK in a way - - after all, he was already famous,
knew tons of key people and could open any door - - - while I was less credentialed and had "a past" that needed
to be closely guarded - - - we were uneasy but incredibly productive collaborators,
each playing our respective roles in doing amazing things we couldn't have done otherwise - -
-
- The OM2: A microprocessor data
path designed using the Mead-Conway methods and layout design
rules
|
|
-
- - - discuss the concepts of cells and subsystems - - - the
new intermediate level of architectural building blocks - - thinking
back to ACS and Sierra - - moving up from gates and F/F's - -
- to subsystems instantiated in 2-D from cells - - -
-
- - - - meantime, Doug Fairbairn and Jim Rowson got the ICARUS
layout system running on the Xerox ALTO workstations - - it was
simple in structure, very fast and quite effective - - it exploited
the newly simplified design methodology - - they created a tutorial
on how it was put together - - it is now clear that their pioneering
work on ICARUS ultimately triggered the evolution of a whole
range of new forms of CAD tools - - setting the EDA industry
off into an entirely new direction - -
-
- But then, what to do with
"the methods"?
-
- - - - then all of a sudden, a remarkable transition occurred
- - - going over and over the simple collection of methods we
now mainly used for most designs - - and bit by bit compressing
the amount of material needed - - - I realized that "wow,
we've got it", i.e., we really had figured out and proven
out in a major design trial an extremely simple, but fully workable,
covering methodology adequate for nMOS digital system design
- - - this was no longer just a dream, a vision that we were
passionately pursuing - - we'd actually got it! - - -
-
- - - -ah, but now we had a completely new kind of problem
- - - what the heck could we do with all this stuff - - this
tiny group of collaborators at PARC and Caltech - - - should
we write papers? - - - but who'd ever publish them? - -
-
- - - - I mentally wrestled with this frustrating problem -
- I could see the reactions to any claims we might make - - after
all, "who were we" - - - or more particularly, "who
was Lynn Conway, this woman who only had an MSEE and who no one
had ever heard of before, to be telling the industry how to do
digital electronics design?" - - - the situation required
that we come up with some "very strong methods" for
even getting our results in front of folks for widespread evaluation
- - much less into acceptance, and use - - - I reflected again
on the past and looked for analogous situations - - and recalled
Steinmetz' propagation of the symbolic methods in courses at
Union College - - I thought a lot about what had happened there
- - -
-
- - - - and then I thought about Dr. Benjamin's 1966 paradigm-shifting
medical textbook - - - and how he had packaged a collection of
avant-garde but serious research results not into scattered journal
papers, but into a comprehensive, compelling, tutorial textbook
that had changed lots of minds about his ideas - - -
-
- A death in the family
-
- - - - in the middle of all the growing intellectual excitement
at work, I got an urgent phone call from my brother - - our mother
had suffered a sudden relapse of the lymphoma cancer she had
thought she'd licked years before - - she was in the hospital
in Austin, Texas, and wanted to see me - - he asked if I could
come down for a day to help cheer her up - - -
-
- - - - I'd seen my mother only a couple of times since my
transition - - - our relations were very strained, because she'd
never accepted my gender transformation - - - she just wouldn't
let go of her "son" and embrace her daughter - - -
no matter how much my physical appearance and outward personality
had morphed, she still managed to see the "old" me
in there somewhere, and she still made mistakes with pronouns
and with my first name - - - however, I went to Austin anyways,
out of concern for someone who was now possibly in deep trouble
- -
-
- - - - my brother met me at the airport and we went to the
hospital together - - - it was a tearful reunion, at least for
for my mother - - - she didn't look well, and she kept saying
"I'm frightened", over and over again - - - there wasn't
much I could say, though I tried to calm her - - - it quickly
became clear that this was not a cheer-up visit; this was to
be the last time I'd see her alive - -
-
- Lynn in 1977
-
-
- - - - at one point she began looking at me very carefully
- - right in the face and eyes - - I was a bit surprised, because
she would never really look right at me during our earlier meetings
- - then after a while, she said "you're beautiful, so beautiful",
with kind of a quizzical, questioning feeling to it - - - I wondered
what the heck she meant by that! - - - was she finally waking
up to the fact that something truly profound had happened to
me nine years ago? - - - and that it wasn't realistic to keep
thinking that she had two sons anymore, but had a son and a daughter
instead? - - who knows - - -
-
- - - - after I'd been there quite a while, I noticed there
hadn't been any other visitors - - that seemed strange since
we had lots of relatives living in the area - - - I asked my
brother about it and found out why - - - my mother had warned
them all that this was the day "I" would be there,
so that they could avoid the terrible embarrassment of "having
to see Robert dressed as a woman" - - -
-
- - - - I'm not sure what they expected would embarrass them
so - - it reminded me of the way the senior execs at IBM had
reacted - - - namely, to an imagined stereotype, rather than
to a real human being who was actually a rather nice person if
you took the time to get to know her - - -
-
- - - when it came time to leave and say our goodbye's - -
my mother said "please, please stay, I'm so afraid"
- - I said that I was sorry, but I had to go - - - that I had
a plane to catch - - I told a small white lie, saying that maybe
I could fly back again soon - - but of course that wouldn't happen,
since all the others would be there till the end came - - - as
it did a few days later - - - and I wasn't invited to the funeral
.
-
- - - - I was so used to such treatment by this time that none
of it caused any emotional stir in me - - no shame, no hurt,
no anger, no tears, no nothing - - - it was as if I were watching
a familiar movie scene - - - the unfolding scene became just
more "ethnographic data" to be filed and referenced
at a later time - - - yet another record of oft-encountered "transphobic"
behaviors - - - I felt like an observer of events that I wasn't
even personally involved in, and was just going through the motions
- - -
-
- - - - This notion that "It's all just data" - -
- that all these behaviors and interactions were useful things
to document as ethnographic data about gender transitions - -
- made events that would otherwise felt very hurtful seem to
be "interesting observations" instead - - - there was
no need to assign blame for what happened one way or the other
- - - these were all just "natural side-effects of a gender
transition" - - -
-
- - - - this way of looking at things came from my studies
in anthropology - - - I began to feel that maybe all these "observations"
would prove useful someday - - - in the meantime, this mental
trick proved very useful in easing the natural pain I felt in
such situations - - -
-
- - - - Anyways, I'd already lost my mother when I was four
years old - - -
-
-
- The idea of writing "the
book"
-
- - - - in a meeting at PARC with Carver, Doug, Jim and few
other folks in the late spring of '77, it was late in the day;
we were tired, and kidding around and winding things down - -
I had finally hit on a specific idea, and I just said it out
loud: "Let's write a book on the new methods!" - -
- "a book that looks like the kind of textbook you might
see after such methods had been used for many years and were
all proven out in practice - - full of design examples, etc."
- - to my surprise, Carver instantly said "Yeah!",
very, very loudly - - and that was it: we set off to do "The
Book" - - thus another of "Lynn's wild projects"
was about to take off - - -
-
- - - during the summer of '77 and on into early '78 Lynn,
Carver, Doug, Jim, Bob
Sproull, Dick
Lyon, and Carlo
Sequin all teamed up to help brainstorm about, create and
test this book - - this was an incredibly talented set of folks,
and we really infected each other with our mutual excitement
about the work - - - a series of versions were written and tried
in preliminary courses - - the first three chapters in a course
taught by Carlo at U.C. Berkeley - - then the first 5 chapters
(including the OM-2 design examples) in a course by Bob Sproull
at CMU - - - at this point we decided to change the title from
Introduction to LSI Systems to Introduction to VLSI
Systems - - just about that time I'd seen "VLSI"
as an acronym a couple of times in Electronics Magazine,
and thought it would add the right additional flair to the title
- - we still called my PARC group the LSI System Area for a while,
but we'd change that a bit later too - -
-
- - - I'd recruited Dick Lyon by then, and he began collaborating
with Bob Sproull to formally define the layout design interchange
format that came to be known as "CIF2.0" - - an extremely
critical element in establishing linkages among all the various
design and fabrication tools - - Dick is a brilliant and creative
researcher, and would go on to invent the optical mouse chip,
invent a complete bit-serial VLSI design methodology for signal
processing applications, extend the scalable design rules to
CMOS, etc., etc. - -
-
- - - we'd also recruited Barbara Baird, who had been the secretary
for the 7100 team at Memorex, to be the administrative assistant
in the new LSI Systems Area - - Barb would prove to be an absolutely
key member of the team during the critical periods ahead - -
-
- Our secret weapons: the Alto,
Ethernet, Laser Printers and Arpanet
-
- - - - our secret weapons in rapidly writing and evolving
the textbook were the Alto computers, laser printers, and Arpanet
access at PARC - - - I did most of the writing and editing of
the emerging text on my Alto - - receiving input from Carver
at Caltech re the OM-2 - - and exchanging drafts, doing editing,
and collaborating with an ever increasing number of contacts
at other places, using e-mail and file transfers via the Arpanet
- - this infrastructure: the Alto personal computers, ethernet
and laser printers at PARC, Arpanet, e-mail, etc., way back in
'77, was very similar in effectiveness to the modern PC's - -
and gave us an amazing ability to rapidly create, distribute
for checking, refine and self-publish drafts of the emerging
text - -
-
- - - - not only that, but the Alto's and the networks presented
a whole new opportunity - - we could organize a "VLSI team"
of collaborators from many different sites - - I'd greatly admired
the effectiveness of CSL's organizational methods (CSL was the
other PARC computer lab) - - - and CSL Lab Manager Bob Taylor's
amazing leadership ability - - - so I somewhat copied the "dealer"
style of CSL - - where folks collaborate and compete in a complex
"open participation" environment, where what counts
is getting something you're working on "get into" the
emerging technology system - - only our "VLSI dealer"
was realized via interactions out in the Arpanet - - on only
a modest scale at first - - but it quickly got bigger, and bigger
- - -
-
- - - there was also a strong techno-ideological component
to our work - - partly a version of the PARC ideology of wanting
to "bring computing to the people" - - everyone would
have their own - - here in our case, it was the notion of wanting
architects and designers to be brought out from under the dominance
of the fabrication technology and facilities - - make them "authors
and writers" who got their stuff "printed in silicon"
- - gaining visibility for the intellectual work of the architect
and designer - - heck, up to now, it seemed that most people
thought that the fab line "made the chip", i.e., was
the source of the intelligence in it - - that's like thinking
that the printing press writes the news articles!- - -
-
- - - - then too, Carver had caught entrepreneur's fever -
- he was always looking for opportunities to start companies
- - - it was becoming clear that a large space of possible start-ups
might be opening up here - - -
-
- - - all these dimensions of thought: our shared vision in
a tight knit team, our new methodological ideas, the wild new
infrastructure we could exploit, our interactions with a growing
team of outside collaborators, the entrepreneurial opportunities
- - all this became intellectually intoxicating - - -
-
- - - - so it was inevitable that we would proceed - - but
we were no longer in charge of this phenomenon, it was in charge
of us - - and it wasn't until years later that I would realize
the personal price I was to pay, physically and emotionally,
for 6 to 8 years lost to anything else but my career - - -
-
- The crash effort
to evolve and propagate the new methods
-
- - - - and thus began a crazy, intense period - - - "this
was it" - - the time to really crank up, and make hay -
- it was now or never - - - and most everything else in my life
ceased, except work - - -
-
- - - - for the next several years, I worked six or seven
days of the week, often for 12 to 16 hours a day - - at my Alto
- - writing, e-mailing, FTPing files, - - - getting up early,
drinking coffee all day - - going home late - - drinking some
wine to finally crash at night - - then back at it the next day
- - - day after day, month after month - - - (by 1980, this phase
had almost ruined my health) - - writing "the book"
- - - and coordinating activities in an ever-enlarging "VLSI
community" out in the network - - -
-
- - - in the spring of '78, Bert suggested an exciting, but
very challenging possibility - - he was on the visiting committee
of the EECS department at MIT - - he'd talked about the Mead-Conway
work with folks there - - he offered me a "sabbatical",
a chance to teach at MIT that next fall and introduce the new
methods there - -
-
- - - - I was very, very intimidated by this prospect due to
lingering shyness with groups of strangers, frequent difficulties
in public speaking, and concerns about being accepted by folks
at M.I.T. sight unseen - - plus my thoughts of M.I.T. were tinged
with deep emotions and powerful memories of my earlier transitional
experiences and romantic affairs while there - - - maybe I feared
that somehow I'd be "outed" there - - - who knows -
- -
-
- - - - but Bert finally convinced me that this was a truly
significant opportunity that could not be missed - - I realized
that this was the opportunity to create and prototype what could
become the standard modern form of VLSI design course, based
on the emerging Mead-Conway text - - I realized that I just had
to do it, no matter how frightened I was - - - somehow Bert seemed
to know just when to encourage me to take the next, just barely
do-able, adventurous step - -
-
- - - we made a crash effort that spring and summer to finish
a complete nine chapter draft of the text - - I put together
Chapter 6, Chuck Seitz
contributed Chapter 7, Chapter 8 was contributed by H.T.
Kung and Carver wrote up Chapter 9 - - we printed up and
bound many copies at PARC, including color plates run on the
new Xerox color copier at PARC - - wow, this thing was starting
to look pretty amazing! - -
-
- - - we'd recruited Alan Bell from BBN to join the LSI Systems
Area - - Alan was a very shy, but extremely brilliant computer
scientist - - we had a hard time convincing him to come to PARC
- - it was a big move for him from familiar territory - - but
he did, and immediately began to play a critical role in prototyping
and evolving infrastructure for supporting and coordinating the
work to be done at both MIT and at PARC regarding project implementation
- - -
-
- - - - our work had also come to the attention of Merrill
Brooksby, manager of CAD development at HP - - Merrill began
to help us coordinate explorations within HP for obtaining QTA
fab - - we got that all set for the fall, with HP benefiting
by being able to "look over our shoulders at what we were
doing" - - -
-
- - - - in mid August, 1978, I set out in my station wagon
from California, headed for M.I.T. in Cambridge, MA, loaded up
with boxes full of books, notes and equipment - - - (by the way,
I kept my 240Z till '83 - - using the station wagon for work,
and the 240Z for play!) - - -
-
- - - - I had a lot to think about during that drive - - here
I was returning after almost 20 years to my old M.I.T environment
where I'd tried to transition so long ago - - - but this time
I was coming as a woman and a faculty member - - - it was almost
too amazing - - - but it was scary too - - - there were many
questions - - - could I really do this? - - - Bert Sutherland
thought I could and that I had to do it - - - but I had in many
ways lived a kind-of insular and protected life at PARC - - always
among the same folks - - - and I was still very shy speaking
in front of groups - - - how could I get up the nerve to teach
and do it several times a week all semester? - - - and in a totally
new environment - - - I was also becoming increasingly worried
that my wider exposure in the research community would lead to
me being "outed" by somebody - - - and that I'd suddenly
lose everything I'd worked for all these years - - -
-
- - - - the saving thing was that I was "on a mission"
- - - the mission of proving-out the Mead-Conway methods and
the draft textbook to be published the next year - - - and once
launched on such a mission, fear wasn't likely to stop me - -
- so off I headed to M.I.T. - - -
-
-
-
-
-
-
-
