Inkwell: Authors and Artists

    

Joining us today is writer Ellen Ullman.

Ellen started working as a computer programmer in the late 1970s.
Twenty years later, she began writing about her experiences in the
industry, describing the effects of technology on both its practitioners, 
the programmers and software engineers, and on society at large.

She is the author of the 1997 memoir _Close to the Machine_ , winner of a
Salon Book Award, and of the recent novel _The Bug_, which was selected by
The New York Times as a Notable Book for 2003. She is a frequent
contributor to Harper's and Salon, and is currently a contributing editor 
at The American Scholar, where she writes the "Scientific Method" column 
each autumn.

Our own Laura Miller will be leading the conversation. Laura Miller is a
journalist and critic living in New York. She is a co-founder of Salon.com, 
where she is currently a staff writer, and a columnist for the New York 
Times Book Review. Her work has appeared in the New York Times, Time 
Magazine, the Wall Street Journal, Entertainment Weekly and other 
publications. She is the editor of the "The Salon.com Reader's Guide
to Contemporary Authors" (Penguin, 2000).

Welcome, Ellen and Laura! Pull up a chair and make yourself comfy.
  

    
Thanks, Cynthia.

Since I'm a literary critic, I'm mostly going to asking Ellen
questions about writing, but if anyone wants to leap in to ask her
about programming, please feel free at any time. If we all work
together, we can probably cover all the aspects of our Renaissance
woman guest.

Ellen, this is the first fiction you've published after making a name
for yourself as an essayist. Lots of writers work in  both modes, of
course, but it's always struck me as a difficult transition, going from
just, well, saying what you think, to orchestrating an elaborate
illusion and making it convincing. What were some of the challenges for
you in doing that. Did you have any other writers you used as models?
  

    
Oh, and sorry for the typos, guys. I forgot you can't edit this stuff
after it's posted.
  

    
I had a difficult transition, Laura, but not for the reason you
mention. I never found essay-writing to be a matter of "just saying
what you think." 

When working on an essay, I always have the feeling that it has to be
a story, that the piece has to show the unfolding of a thought, how I
came to the ideas, and how I feel about those ideas, not just the final
"here is what I have concluded." I'm writing now for The American
Scholar, and this is giving me a chance to further explore the notion
of the essay as a thought-story, a sensibility *moving* over a
collection of thoughts to come to some conclusion,which itself may be
temporary, since life of course keeps moving.

That said, I had an awful time writing The Bug after Close To Then
Machine. I had to give up the authority of the this-really-happened.
And it took me years to find a fictional voice that had appeared to
have the authority that a nonfiction voice naturally has. 

My model for the essay has always been Cythia Ozick. She is
passionately smart. Or intellectually passionate. Or any way it's
possible to express the sort of mental intensity that becomes emotional
by nature of the sheer force it exerts.

For fiction, my models have always been spare writers -- Kafka,
Sebald, Coetzee -- thought I myself, to my horror, am not really a
spare writer. There is a big problem when your taste doesn't match your
own inclinations. Finally I just had to let go and let the story be
what it would be, let myself speak in the voice that was most fluent,
for better or for worse.
  

    
There's also creating the characters. I've heard you say that most
readers just assume that Roberta, the tester character in "The Bug," is
based on you, while you feel that Ethan is a closer match.

Probably they're making the assumption that a female character must be
the avatar of the female author, but I can understand the mistake all
the same. Roberta like you, is a person with literary and cultural
interests, while Ethan is almost blinkered, focused on his work and his
artificial life project. I wonder if they're both not a version of
you, often in serious conflict
  

    
Yes, you have it right: they're both me. And yes, often in serious
conflict. 

The literary and engineering impulses are at odds -- intrinsically, I
think. I mean, what moves me as a engineer is the desire to get
something to *work*. It's full of fiddling and exploration, like
writing, but ultimately engineering's pleasure is to create something
functional. 

Fiction, on the other hand, is purposeless, in the best sense. The
only way it "works" is in communicating the interior experience of
being alive. 

(That is, when it does indeed work, which of course is for the reader
to decide.)
  

    
Some of the scenes that most stick out in my memory of the novel are
meeting scenes at the software company where Roberta and Ethan work. In
particular, there's a great one where the venture capitalists who are
funding the project have come in to get the deadline moved up and Ethan
is called into his bosses office to talk about it. 

The only place to sit is one of those big bouncy inflatable exercise
balls, and Ethan has to struggle to keep his balance on it while
sweating out the meeting (he's not good with people and he starts to
worry that he'll be caught in some of his resume exaggerations, or say
something that the vc's won't like).

Can you tell us a bit about how that scene came to be?
  

    
That scene is completely made up. It's a melange of things: my
experiences with venture capitalists, exercise balls and toys I'd seen
on a visit to Google in its early days, meetings I'd had with a boss
for whom I worked at one time, my own experiences with the absurdities
of software scheduling. All I knew when I sat down to write the scene
was that Ethan was someone motivated by the fear of humiliation, and
that he had to be struggling against appearing ridiculous. 

It's odd. That scene is one of the few places in the book where I
allowed myself to simply let things go and see how silly a scene could
get before it got annoying. It felt risky while I wrote it. But maybe I
should be doing more of that (it occurs to me now).
  

    
I was gonna say, as soon as I saw the exercise ball -- Google.

Ellen, how the heck are you?  It's been years. Are you still in the
city?  I'm spacing out now on the book you wrote a few years back that
got so much attention, but what sort of effect has that had on your
life since? 

and how's your kitty?
  

    
(note: offsite readers can join the conversation by sending their questions
or comments to <inkwell-hosts@well.com> )
  

    
Ellen, welcome. I have to say that I really, really enjoyed reading "The 
Bug." I think that, to some degree, I liked it more than the essays in 
"Close to the machine." Often, in the engagement with how working with 
technology affects (or doesn't affect) the way we are, who we are, I found 
myself nodding the way I do when reading a particularly good Richard 
Powers novel ("Plowing the Dark," for example).

In that regard, I especially value that the book seems to be about 
technology and people who work with technology, as much as it is a dramatic 
story. 

Having said that, the last chapter or two felt completely anticlimactic. 
It's as though once we know the fate of one of the main characters the 
story was over. You went through the motions of wrapping loose ends up 
neatly, but the point was what happened to that character, and then, I 
dunno what happened? Am I the only person who feels that the novel ended 
and you kept writing? How did you feel about the ending?
  

    
Sorry to skip the topic around, but I forgot to mention one
particularly nice detail about that scene wtih the toy-filled office,
how most of the toys have an aggressive purpose, even as their playing
at being playful, you might say. it's a squirt-gun, one of those
suction-cup bow and arrow sets, etc. The one toy that might seem
innocently exhuberant, a mechanical bear that claps when it's wound up,
gets ritually wound up by someone who's made a mistake.

I didn't even notice this the first time around, but it v. deftly
undermines the menace in that workplace, the way the people seem
divided against each other and dangerous to each other despite the
rompish facade.
  

    
>How did I feel about the end?

Well, I did feel I had to say how the bug was solved. Otherwise it
would be like a murder mystery that ends with, "And we never found out
who did it." 
  

    
>the way people seem divided against each other

Yes, I did mean the wind-up toy to be menancing. The scene ends with
Ethan and Harry staring at it as it paws at the air.

I have to say that my experiences in startup companies were full of
pressure, the sort of pressures for success and wealth that, to me,
felt like menace.
  

    
years ago, a short time after I had been promoted from programmer to
project leader, I was griping to a friend about all the interpersonal
stuff that I had to deal with.

"Ahhh, the human element" was his reply.

It only gets worse when there's the sense that there's a pile of money 
to be made.  

Take me back to the code.
  

    
I agree. The code becomes a way to dodge the messy, unfriendly
relationships. But of course this only exacerbates the problem. 

There is this idea that programming is a logical process, and that the
human interaction is somehow a distraction from the orderliness of the
code. But there is nothing more political than a software development
project. There are product managers, investors, potential clients,
existing clients, technical people, salespeople all pulling for their
own approaches. The process is full of concessions to "special interest
groups." In contrast, the passage of a law in Congress looks
positively mathematical.
  

    
I used to give a speech on the five jobs of the project manager.  The 
idea was to show that the only way to keep the job was to focus on 
keeping the project funded by the sponsor and delegate everything else, 
then work on the biggest messes until you get it right.  

It is kind of amazing that any software actually works if it was 
designed and written by committee.  
  

    


Fin Hunt writes:


It sounds like a terrific book. I'm going to Barnes And Noble to check it
out. There was a book that came out a few years ago, I forget the exact
title, "Inside Apple" or something, that talked about Apple's rise and fall
and resurrection. It was great. Fascinating to read about all the things that
might have been. For each piece of software that comes out, there's a lot of
hardship involved.
  

    
>It is kind of amazing that any software actually works if it was 
designed and written by committee.  

It constantly amazes me. Of course, it doesn't work very well. Hence,
The Bug.
  

    
Artificial life is another theme in the book, this hobby that Ethan
works on when life itself gets to be more than he can take. Did you
ever dabble in it yourself, Ellen? What's the attraction?
  

    
Howdy, Ellen, I really enjoyed reading The Bug.  A mystery with a
logical conclusion. 

I never worked at start-ups, rather I was a corporate drone. One thing
that kept hitting me was the emphasis on graduate studies in this
book, and the appraising of people by their educational background
rather than their actual work experience. Is that common in start-ups?
  

    
No, I don't think it's common for start-ups to put much emphasis on
educational backgrounds. On the contrary, I think part of the whole
dot-com illusion was that everyone was supposedly a brilliant drop-out,
a la Bill Gates or Larry Page (of google, who never finished his
Ph.D).

The Bug is fiction, and the story is particular to the characters. In
this case, having to leave graduate school was a defining moment in
Ethan Levin's life. He had to pick up programming as he went along, and
this fed his sense of insecurity. Then he had a tester who was far
better educated than he was, and so the conflict is put in place.

That said, I based some of The Bug on my experience at Sybase, where
the woman who set up the technical support, testing, and documentation
departments (everything but original engineering) had a Ph.D. This was
the mid-1980s. Unable to find people with experience, she decided she'd
hire smart people who would learn, and so hired all these people with
Ph.D.s in things like classics and linguistics. We had all these
programmers who had BSEEs, and their testers were wildly
better-educated than they were. It made for interesting relations, as
you might imagine.

There is an inherent antagonism between programmers and testers. A
tester is someone whose whole purpose at work is to find out everything
you did wrong. Add to that the difference in temperment between
engineers and people who read three ancient languages, and things get
interesting.
  

    
Now to Laura's question about artificial life:

When I was starting out as a programmer, in the late 1970s, early
artificial intelligence research was still at a peak. It was a time of
great (and naive) optimism, when computer scientists like Marvin Minsky
believed that an intelligent machine could be devised within a decade
or so. I poured over the literature but found myself skeptical. My
books from those days, which I've kept, are full of ridiculing marginal
comments like "Oh yeah?" and "You've got to be kidding!"

Many years later I had occasion to get to know Chris Langton, who is
one of the important figures in artificial life. Chris is known for his
work at the Sante Fe Institute, where he championed the idea of
intelligence as something that emerges from collections of low-level,
"dumb" interactions. He created so-called self-replicating figures
known as Langton's loops, which propagate themselves and evolve by
following simple, local rules.

Chris told me that when his loops first evolved, he sat there and
cried. I never forgot that. The image of a man crying before his
evolving digital creatures stayed with me, and I knew that somewhere in
this story, Ethan Levin would cry in front of his creations. I didn't
know the exact circumstances -- I didn't know what the emotional color
would be, what exactly would be happening before and afterward -- but I
knew that something like that scene would be in there.
  

    
Fascinating. But given that cultural difference, how common is it for
testers to leap over the wall, as it were, and get into programming?
  

    
Well, Roberta is more or less pushed over the wall. As a result of
Ethan's complaints about his work being tested by someone who does't
even know how to program, she gets assigned to take a class in it. She
gets sucked in pretty fast, and she gets some sense of what it's like
to be Ethan, that is, buffered from a more direct experience of the
world by this intellectual preoccupation that's always running in the
back of his mind.

Ellen, you write so effectively about the allure of this kind of work
and the satisfactions it offers, but you also seem ambivalent about it.
Do you still think the work demands the sort of autistic culture that
tends to grow up around it, or is that a tendancy that can be overcome?
  

    
Roberta is pushed into it, as are most testers, but only to a limited
extent. I've never been a tester, but I do cringe when I remember how I
felt about those who were not technically able. It took me a while to
understand that software is not for the programmer but for the *user*,
and that the tester represents the user's experience.

As computers have advanced, and more of the processing power is being
used for the UI, there are many more people involved in software
development than just programmers. A whole new specialty has been
created -- with titles like "Director of User Experience" -- in which
nitty-gritty knowledge of code is not important but knowing how to
figure out users' reactions is. These new UE specialists are, believe
it or not, Ph.D.s in things like anthropology -- which describes
Roberta's boss, Wallis.

I do think that the profession of programming is changing. Programmers
may *want* to retreat into a semi-autistic world, but so much of
coding these days involves the use of pre-existing code, packages, and
libraries, that it's nearly impossible to just sit by yourself and
write programs. Open-source development has made programming more of a
social act: your code is seen by hundreds if not thousands of other
programmers, each of whom can have a very public reaction.

That said, the sort of society that has been created in the
open-source world is not exactly a tea party. I'll probably get a lot
of flack for saying this, but the people involved don't always have the
best people-skills. The interactions are very rough and tumble,
tending toward a subtext of who is smarter and better and right.

For example, after The Bug came out there was a discussion on
slashdot, the programmer site. One guy, complaining about something he
thought was wrong in the book, said something like, "I guess my
twenty-five years of programming trumps her [my] twenty years." And
then someone replied to *him*, saying, no, he was wrong, and "I guess
my thirty years trumps your twenty-five." 

If I had written that interchange in the book (which I easily might
have, since it seemed so usual to me), I would have been criticized for
being anti-technical-- as I have been. Programmers seem to want to
have books that reflect their experience, but many simply hate to see
anything negative said about it. I understand that. There have been so
many ridiculous things said about so-called nerds. So much of it is
uninformed. But like all stereotypes, there is a grain of truth in the
nerd image.

I think, Laura, that there is something inherent in programming that
will always attract a certain sort of person. After all, most of your
working day involves interaction with a compiler, a very dumb,
literal-minded thing. Anyone who is gregarious or dreamy or
free-associative in their thinking -- someone with a metaphorical
imagination -- simply wouldn't be able withstand the sort
single-mindedness that programming requires.

And I think we are stuck with this. After all, the computer was built
to take care of the things we're bad at -- literal-mindedness,
recalling long lists, doing repetitive calculations. We can't then
expect it to be dreamy and poetic. What we can do is make it *appear*
to be dreamy and poetic. But underneath it all, someone will still have
to write the code that creates the illusion.