Women and CompSci: What does it mean for DH?
Ben Brumfield
group already, but between incompleteness and illegibility, it's
pretty awful. Here's my recap of the discussion, with a bit of
elaboration and expansion of abbreviations:
It's a bit hard for me to reconstruct, but the discussion started when
someone pointed out the male-female ratio among Dan Cohen's twitter
followers, using TwitterAnalyzer.com. Various people (including
Melissa Terras, Sterling Fluharty, and others I can't recover) started
looking at the ratio among Amanda French's followers and discussing
the methodology the tool uses. Pretty rapidly, however, the
discussion took a turn towards the implications of the finding that
between 60% and 70% of DH followers are male: what does that say about
Digital Humanities?
As a software developer, I'm no stranger to such debates. The paucity
of women in CS programs and IT in general is notorious. Every few
years, the subject rises within programming communities and provokes a
wave of agonized soul-searching. Typically a prominent programmer will
write or say something transparently misogynist or tactless--most
recently Matt Aimonetti's "Perform Like a Porn Star" presentation at a
Ruby on Rails conference--whereupon other programmers will point a
finger at him, saying "This kind of $%& is why there are no women in
programming!" At that point, folks look around at the single-digit
proportion of women in their community and the internet fills with the
noise of a predictable, three-sided argument. One side comprises the
minority who believe that overt sexism explains why so few women are
in Computer Science, another side is the misogynist minority who
actually think that women can't code, but the majority alternates
between agonized self-flagellation and defensiveness: the simplistic
explanation that overt sexism is what keeps women out of CS implicates
themselves in that sexism. And not sexism in the "patriarchy is a
system that harms us all" sense, but in that universally-anathematized
1950s sense. Sadly, a few of that bewildered majority preserves their
self-image by joining the "women can't code" crowd.
As is probably obvious by now, this quarrel among software developers
is essentially context-free -- wholly uninformed by the decades of
research into women (and girls!) in CS, science, or engineering. As a
result, the participants take the arguments very personally, and
tempers rise. When Sara Brumfield ran a lightening talk (http://
tinyurl.com/nwnp5v) on the subject at Lone Star Ruby Conference last
year, introducing some basic history and theory, one person commented
that it was the first discussion of women in CS he'd ever attended
"where no one got mad". Sara introduced the attendees to the pipeline
problem--the steep fall-off in the number of women on the CS track
which occurs in middle and high school--and suggested tapping a
"second pipeline" of women who start programming in adulthood. After
all, the ratios are far better in high-tech positions like support or
consulting, which require no less intelligence or technical savvy than
development but do not require Computer Science degrees. One of her
examples of a second pipeline was the Digital Humanities.
With that background, I was astonished to see Shane Landrum's comment
that "As much as I'm excited about DH, I have qualms about its
tendency to reinforce existing pipeline-narrowing effects for people
of color and women. Combining the weed-out effects of computer science
and of humanities graduate programs leaves digital history more of a
white boys' club than it should be." Upon reflection, however, it
makes sense -- Shane can explain his perspective better than I can,
but if women's representation within DH is lower than within the
analog humanities, is the rise of digital humanities a step towards
the kind of lopsided ratios you find within software development? As
Shane elaborated, "retracking women into CS after the middle-school
contraction is great, but building a subfield that requires CS skills
*plus* post-BA-level
humanities means building a (sub)field open mostly to the economically
elite."
I responded that "there's a big difference between requiring CS skills
and requiring a CS degree for success. I gather that most DH folks are
tech autodidacts, their formal training being in the humanities.
That's a very good structure for equality. And a good structure for
successful DH projects, which require less hardcore CS and more of
what Dorothea Salo calls 'banging
rocks together.'"
Shane agreed with me in part, but added that "DH will only remain
accessible for tech autodidacts if DHers keep that an explicit goal."
This is a really interesting statement, since at this point we both
are tying gender equity within the Digital Humanities to their
openness to--and indeed reliance upon--self-taught technologists. I
think that this is correct, but I'm less concerned about preserving
this openness than Shane is, since I think that it might be
inescapable for structural reasons. I responded with this:
"What are the conditions under which DH could restrict itself to the
narrow (and male-heavy) CS pipeline?
1. Abandoning the Web as a platform. This is unlikely, and so long as
HTML markup is the entry-point for the DH pipeline, prospects for
gender proportionality are good.
2. Serious dependence on CS theory, rather than OSS tools. This is
also unlikely, since A) Most DH is what industry calls integration
work, rather than pure development, and that's a good thing for
autodidacts, and B) cutting-edge DH projects != cutting edge CS
research/theory -- show me a successful (measured by running code)
collaboration
by an academic CS department on a humanities project.
3. Scads of unemployed CS grads who turn to well-funded academia-based
DH projects for work, allowing DH to be more selective.
I don't see those happening soon. Have I missed anything? (Like grant
requirements or something?) "
Perhaps this is complacency on my part--I'll admit that most
commercial and F/OSS developers I know would be delighted if we were
to achieve even a 30% F/M ratio--but I see Digital Humanities as a
positive force for women in technology.
Ben Brumfield
Shane Landrum
(What follows is a long reply to a long post. Replies would do well to
split it up topically with new subject headings.)
On Wed, Jan 20, 2010 at 8:47 AM, Ben Brumfield <benw...@gmail.com> wrote:
[snip]
>
> It's a bit hard for me to reconstruct, but the discussion started when
> someone pointed out the male-female ratio among Dan Cohen's twitter
> followers, using TwitterAnalyzer.com. Various people (including
> Melissa Terras, Sterling Fluharty, and others I can't recover) started
> looking at the ratio among Amanda French's followers and discussing
> the methodology the tool uses. Pretty rapidly, however, the
> discussion took a turn towards the implications of the finding that
> between 60% and 70% of DH followers are male: what does that say about
> Digital Humanities?
And I have to say that I'm a little surprised it's only a 70% ratio.
Most pure-software endeavors I've ever seen tilt towards 80-90% male.
> As a software developer, I'm no stranger to such debates. The paucity
> of women in CS programs and IT in general is notorious. Every few
> years, the subject rises within programming communities and provokes a
> wave of agonized soul-searching. Typically a prominent programmer will
"Every few years?" Maybe that's how it appears to some, but I've been
following such discussions for at least the last 10. My introduction
to open source software was Kirrily Robert's website infotrope.net,
sometime in the late 1990s, and she's been a near-constant voice for
women in open source since then. (For those reading this who are new
to the context Ben mentions, I'd suggest a ramble through
http://geekfeminism.org , which is a group blog on women and
technology, broadly defined.)
Sara Brumfield's talk at LSRC is in the spirit of those conversations,
though I hadn't read it until Ben pointed to it.I'm not a close
follower of Ruby as a language, but Ruby has a reputation as one of
the more women-friendly developer communities. It's quite possible
that DH can provide a "second pipeline" into computer-intensive work
for girls and women who've fallen out of the primary pipeline around
middle school. That's great, but it doesn't fix the problem of the
middle-school contraction, let alone digital-divide effects around
social class and access to computer training. (Which is another topic
for another thread.)
> With that background, I was astonished to see Shane Landrum's comment
> that "As much as I'm excited about DH, I have qualms about its
> tendency to reinforce existing pipeline-narrowing effects for people
> of color and women. Combining the weed-out effects of computer science
> and of humanities graduate programs leaves digital history more of a
> white boys' club than it should be." Upon reflection, however, it
> makes sense -- Shane can explain his perspective better than I can,
> but if women's representation within DH is lower than within the
> analog humanities, is the rise of digital humanities a step towards
> the kind of lopsided ratios you find within software development? As
> Shane elaborated, "retracking women into CS after the middle-school
> contraction is great, but building a subfield that requires CS skills
> *plus* post-BA-level
> humanities means building a (sub)field open mostly to the economically
> elite."
So, DH is an emerging field/interdiscipline, and everything I say here
is based on my own, autodidactic, sense of it as it emerges. I'm a
historian by graduate training, but I have an undergrad degree in
computer science and American studies, and I worked for 6 years as a
software engineer before entering my Ph.D. program. Here's what I see,
as far as gendered barriers.
-The ability to know your (software) tools and rewire them to do what
you want is key to a certain level of proficiency with code. Usually,
this degree of proficiency requires a combination of sustained focus
over time, and somewhere to get your newbie questions answered. (The
Hacker Jargon File refers to this as "larval stage":
http://www.outpost9.com/reference/jargon/jargon_27.html#TAG1011 )
- Many segments of hacker culture, including open source software,
have a history of alienating women by tolerating antisocial and/or
sexist behavior. They trumpet technical superiority as the measure of
whether someone's contribution is "good" or "worthwhile," and are
relatively prone to macho posturing behaviors unless there's a vocal
community leader to keep them in check.
- To the extent that aspiring DH people get their skills
autodidactically and rely on open-source software, they're likely to
run into these cultural bugs. (I could tell a longer story about the
IRC #perl channels I used to frequent, but suffice to say that
building a non-woman-hostile environment there required constant
solidarity among female hackers and their allies.)
- And then there's the issue of unstructured time and women's lives.
People who study pipeline-narrowing effects have argued that many
girls/women are socialized to avoid overspecialization or "uncool"
levels of expertise, and they've placed that pipeline-narrowing
sometime around ages 10-14. By the time someone's approaching the
"second pipeline" of DH, she's likely to be older and may have other
commitments that make it harder to carve out time for larval-stage
focus. Meanwhile, her male peers who never fell out of the first
pipeline are getting their degrees in CS and going on to high-paying
jobs in private industry.
Anecdotally, as Ben reports, "most DH folks are tech autodidacts." He
sees this as a good structure for equality, and I agree, but I don't
think the playing field is level. Most of the people I'm aware of who
write code for DH projects are guys, and many women I've talked to
about DH profess frustration at not knowing where to start to acquire
the more code-heavy skills. That was the context for this:
> Shane agreed with me in part, but added that "DH will only remain
> accessible for tech autodidacts if DHers keep that an explicit goal."
I'm referring to structural factors. DH relies, in many cases, on
institutional infrastructures; the most common form of these I see at
universities is a Center for Digital Humanities, by whatever name.
(The Maryland Institute for Technology in the Humanities and the
Center for History and New Media come to mind.) These centers file for
grants to start or maintain DH projects, often in combination with
affiliated researchers on their university's faculty. It's one thing
to say that "anyone can acquire the skills," but when autodidacts
start to build infrastructural power, I think it's important to use
that power to build a more diverse group of DH researchers.
> inescapable for structural reasons. I responded with this:
>
> "What are the conditions under which DH could restrict itself to the
> narrow (and male-heavy) CS pipeline?
> 1. Abandoning the Web as a platform. This is unlikely, and so long as
> HTML markup is the entry-point for the DH pipeline, prospects for
> gender proportionality are good.
Yes, but in technical terms, knowing HTML is like knowing how to make
scrambled eggs; useful, but not that difficult and not that
high-paying. The DH pipeline skills are more akin to making quiche,
soufflé, or croissants; even if you can make scrambled eggs, it'll
take a bunch of practice before you can make a chef-quality soufflé.
Besides HTML, the other skills required for a technical person on a DH
web-project team include at least 1 programming language (Javascript,
Python, Perl, Ruby, whatever), some knowledge of command-line
Unix-style systems, and some system administration experience doesn't
hurt-—not to mention desktop or command-line image processing
software. Mapping-related projects require a familiarity with GIS
tools, which aren't part of the HTML tool stack; textual projects
usually have some connection to TEI.
Anyone can learn to make a soufflé, but it's a lot easier to learn if
you've got access to a good chef's notes or cookbook. From my
perspective, DH (as an emerging field) has a few star chefs making
beautiful soufflés, very few cookbooks, and lots of scrambled eggs. I
want it to be a field where the star chefs aren't all white guys, and
I'm concerned that it will tend to be that way unless people who care
do specific work. Precisely what work that is remains to be seen.
> 2. Serious dependence on CS theory, rather than OSS tools. This is
> also unlikely, since A) Most DH is what industry calls integration
> work, rather than pure development, and that's a good thing for
> autodidacts, and B) cutting-edge DH projects != cutting edge CS
> research/theory -- show me a successful (measured by running code)
> collaboration
> by an academic CS department on a humanities project.
Fair point, though I know there's some good work out there in social
network analysis being done by people who apply heavy CS to the
humanities and social sciences. (If anyone reading this has more
specific suggestions, I'd like to see them. This is one of my current
DH interests.)
What I *have* seen over the last 10+ years is the practical end of the
computer autodidact who could make a good living without a bachelor's
degree, and that informs my thoughts about DH. As far as I can tell, a
"digital humanities person" usually has a BA, MA or Ph.D. in a
humanities discipline and some knowledge of (or interest in) applying
computational tools for research. As DH becomes a better-understood
specialty, I think we're likely to see a push for describing what
*specific* technical skills are useful, and I think that could well
work to limit the diversity of the field even further.
Considerations of social class and access to education are
particularly on my mind here. The educational path I've described
above involves adequately college-preparatory K-12 schooling, the
ability to pay for a 4-year bachelor's degree (and availability of
financial aid), possibly 2 to 8 years of graduate school beyond that
(for which one either goes into debt or lives on a small stipend),
plus acquiring some computer-specific skills besides. DH has a long,
expensive career path, and in my mind, this is a significant factor
for the viability and diversity of the field in the long run.
srl
--
Shane Landrum <s...@cliotropic.org>
Ph.D. candidate, American History, Brandeis University
Hugh Cayless
On Jan 20, 2010, at 11:19 AM, Shane Landrum wrote:
> Anecdotally, as Ben reports, "most DH folks are tech autodidacts." He
> sees this as a good structure for equality, and I agree, but I don't
> think the playing field is level. Most of the people I'm aware of who
> write code for DH projects are guys, and many women I've talked to
> about DH profess frustration at not knowing where to start to acquire
> the more code-heavy skills.
I'm a tech autodidact (more or less), and a guy, but my own experience in a decade or so of DH projects, as well as a stint in private industry, is that the DH world (for whatever reasons) has a markedly better gender balance. I've had the good fortune to work with some very accomplished women (in programming and other roles) on several projects. Contrast that to the last company I worked at, where we had a hard time even finding female programmers to interview for jobs.
There's definitely more tolerance for people learning on the job in DH projects—probably due at least in part to tight budgets and the difficulty in finding people with all the requisite skills.
> That was the context for this:
>
>> Shane agreed with me in part, but added that "DH will only remain
>> accessible for tech autodidacts if DHers keep that an explicit goal."
>
> I'm referring to structural factors. DH relies, in many cases, on
> institutional infrastructures; the most common form of these I see at
> universities is a Center for Digital Humanities, by whatever name.
Another important DH player in many institutions is the Library, which tends to employ more women (though not necessarily in technical roles).
> (The Maryland Institute for Technology in the Humanities and the
> Center for History and New Media come to mind.) These centers file for
> grants to start or maintain DH projects, often in combination with
> affiliated researchers on their university's faculty. It's one thing
> to say that "anyone can acquire the skills," but when autodidacts
> start to build infrastructural power, I think it's important to use
> that power to build a more diverse group of DH researchers.
I think DH is now (and I hope it stays this way) far more open than most "disciplines" in the academy. Part of this is due to its operating on the margins in many institutions. I worry a bit that now it's becoming accepted, it will be co-opted by the traditional disciplines and outsiders will be locked out. But that discussion is probably orthogonal to the question of gender balance.
Best,
Hugh
hmprescott
taught folks, and I would say my skills don't even rise to the level
of scrambled eggs!
The reason I've been able to get into this field at all is that I'm a
tenured full professor and have the privilege of being able to take
risks (and well as a desire to do new things so I don't get bored). I
was also able to get generous funding from my university to attend a
workshop at the Center for History and New Media, and will apply for
more funding to further develop my graduate course.
Also, as a historian of women in science, technology, engineering, and
medicine (aka STEM), I would caution against creating a second (and
lesser) track for women in CS and IT. One of my graduate students who
took my digital history courses is an administrative assistant at
another CSU campus. She's getting her M.A. so she can get a better
job and stop getting treated as "just a secretary" by faculty who know
far less about computer technology than she does. The female
librarians also feel like their disrespected by other faculty (at my
university, librarians are in the same collective bargaining unit as
teaching faculty and subject to similar criteria in terms of promotion
and tenure).