GirlsinOpenSource
May 15th, 2008
I do dozens of press and analyst briefings each year and I’m usually pretty confident that I can handle any question that comes my way. I’ve come a long way since that time at Comdex when I declared to an audience that included a lot of press members what I would have chosen as an alternate career. More on that in a future blog.
I did an interview recently the preparation for which had me totally perplexed. I understood from the briefing sheet that the interview would include questions about how a data center run by a woman would differ from one run by a man. I’ve never run a data center, and if I had I don’t think I would have put up window treatments and floral wallpaper, so I turned to my colleagues for input in answering this question. The one consistent answer I received that resonated with me is that women are more collaborative and, in a time of crisis, will pull a team together to get the problem resolved. Since most data centers typically operate smoothly under a standard set of procedures, it’s how things are handled at crisis points that separates the women from the boys.
My interviewer, Jessica Twentyman, had done her homework. The questions she posed to me all related to the role women play in engineering, a topic with which I’m intimately familiar. We had a great chat, and the one question I had to think long and hard about was how code written by a woman would differ from that written by a man, and whether or not I’d be able to identify the gender of the author of a piece of code. This is nothing I’d ever thought about before, and given our strict coding standards at Ingres, our code is fairly androgynous. My answer to that question, as well as the rest of the interview is available online at the Financial Times. I’d love to hear your thoughts about this issue: Do you think there’s a difference between how men and women write code, and if so, how does the code differ?
From my days at CA I know two of the other women interviewed for the Financial Times article, Siki Giunta and Jacqueline de Rojas. Who knew that CA was a breeding ground for successful women in IT?
We have a viral marketing campaign underway at Ingres right now promoting women in open source, and we made some great t-shirts as part of that campaign with this phrase silkscreened across the chest: GirlsinOpenSource. There’s a story behind the t-shirts that I’ll share in a future blog. If you’re a woman in open source we’d love to hear from you, so drop us a line and we’ll start up a conversation—and get a t-shirt out to you in the mail.
May 15th, 2008 at 2:41 pm
I have, personally, not noticed difference due to gender but I have, for sure, do noticed difference due to personalities.
As Emma says we have strict coding standards, but before Ingres I have worked in different companies in France, on different languages but more specifically in C and a bit in 4GL (ABF, OpenROAD), on some projects we were 50%/50% men/women. By looking at the code, I could tell who from my colleagues had written it, but I think the way a code is written, reflects more the personality of somebody than his/her gender. I remember specially a guy who was used to write pages and pages of code, at one time I had to reduce his code by 5 if not more. But that was the way he was, always talking and talking which is for sure not my case :-). So I will say that I have indeed noticed difference due to personalities, but I am not sure about gender.
May 15th, 2008 at 6:06 pm
I’m not sure I can agree with the statement about being able to tell the difference between code written by a woman and code written by a man. It’s obviously difficult to generalize anyway but I think my take on it would be, good code is good code and bad code is awful;-) There may be a difference in the way we (ladies and gents) approach a problem but the end result still needs to be clear, well annotated, reasonably easy to follow and free from too much unnecessary obfuscation.
Having worked with both men and women coding a variety of languages over the years I don’t think I’ve noticed a discernible difference between the two groups or similarity within, but this may just be because I’m a guy
Great article all the same though, kudos.
May 16th, 2008 at 5:19 am
Without even bringing code into the equation (no pun intended), I can say that men generally like to show off. From the time we are little we do wheelies, jump from the roof into the pool (not recommended), and all sorts of things to show how generally impressive we are. And our code is no different. I’m willing to bet recursion was first used by a male :). In most cases, it’s like doing backflips down the street when walking would do just fine. Maybe this could explain the abundance of men involved in open-source projects. There is by nature a larger audience for your backflips. In our defense, I like to believe that this often results in useful discovery. By contrast, the female-written code I have seen tended to be to-the-point, with more comments, and a bit more emphasis on organization. If only these qualities could be combined, our code would be fancy, yet easy to read and modify. Note to self: engineer super-human race of hermaphroditic code monkeys.
May 16th, 2008 at 7:08 am
As I never was a real programmer this question is not so easy to answer for me. I also do not look at too much code apart from small sample code snippets
But I could imagine that women better document what they are doing, whereas men could more put some tricky code which afterwards has to be analyzed from the person who is coming to it.
This is at least the case for the testcases I get
May 16th, 2008 at 9:49 am
There was a BBC programme recently about Bletchley Park.
As we all know the computer was invented there and
was used to shorten World War II. My recollection of
the interviewees was that there was a good mix
but that the women were all involved either in clerical work
or in monitoring enemy communications. It
seemed that the men did all the new fangled computer
stuff.
Today the proportion of women involved in writing
software is still way too low. Maybe it has to do with
history but I am sure that the field would benefit
from a more balanced population. I wonder if the
competitive nature of the business puts the more
collaboratively inclined _people_ off. We all know that
we have to collaborate to succeed, but us men still want to
everyone to know that we create the coolest code (or
perform the best back flips.)
Even so I am far from convinced that it is possible to
tell the author of a piece of code that has been
maintained through a number of cycles and adheres to a
coding standard like the Ingres one which is designed
to promote collaboration, after all
I am convinced that collaboration is a clear
differentiator for Open Source. The Eclipse project is
an example that is very dependent on collaboration.
Eclipse has a “release train” that pulls into the station every
summer: In June 2007 the Eclipse
Foundation released Eclipse Europa with 21 projects and
over 17 million lines of code with contributions from
310 developers in 19 countries. The release was on
schedule. Apart from something to be hugely proud of,
such an
achievement requires a high degree of collaboration.
Alan Turing who played a key role at Bletchley Park was
remembered by one of the interviewees who recalled
that he had some a problem with his bicycle chain
and counted
the number of times the pedals went round so as to stop
before the chain came off. While that isn’t exactly a
back flip, I think it might count?
May 20th, 2008 at 6:54 am
I have always felt that there are far too few women in IT.
I also feel that the differences in coding style between men and women can sometimes can be compared to driving. Men usually polarise between two extremes; some are very technical and quick the rest……..! Women on the other hand have no extremes. They are generally very competent and have significantly less accidents!
I may add that women make excellent team leaders and managers.
June 16th, 2008 at 6:58 am
Do any of you have any empirical evidence for your comments?
Anecdotal != Evidence
My own anecdotal experiences have been that most women programmers are horrible programmers. But that is not evidence that all or even many women are bad programmers.
I have seen a woman with a masters degree in CS that took exceptions returned by the testers and just hardcoded the expected results so that they would pass the test if the same data was used. I ended up debugging all her code and she ended up fixing spelling mistakes in the user interface. She also got good reviews and was promoted into management.
However, I am a man and realize that she was just stupid and that the organization was required by law to promote a certain number of women and since we had very few in the IT field, she gets a free ride. If she had been a minority too, she may have made CEO.
It never occurred to me to generalize about all women based upon this (and a few other similar cases). Maybe all the good ones were working for other companies. I know my sister can code and I have often heard it repeated that she is a girl.
In other words, I think you ladies are being a bit sexist. A man would be castrated for saying anything similar. Just check your response to my post.
June 16th, 2008 at 6:58 am
In truth, this is just a sexist, sensationalist article. The veracity of such a blanket conclusion drawn from a limited pool of skewed “statistics,” is dubious at best.
What others are saying: http://developers.slashdot.org/article.pl?sid=08/06/16/1212240
June 16th, 2008 at 6:58 am
Emma,
I understand and applaud your desire to see a more balanced development workplace and I do believe that mixing diverse modes of thought results in synergy. I would very much like to see more women in software development. However, I had the opportunity to read your recent WSJ interview (which may or may not be the same interview you referred to in this post; I can’t check the FT interview without registering on their site) and I have to disagree with your conclusions that women tend to document their code better than men and that men tend to obfuscate deliberately to show off.
First, there was a scientific study performed on this very topic that refutes your claims. This study found that the factors you mention had poor discriminative value in classification of gender, and further, that subjects were usually unable to distinguish code written by men from that written by women. It seems more likely to me that the perceived ability to distinguish between the two is more likely the result of an unbalanced class distribution (80% male, 20% female). If you were to attempt to apply this distinction in a balanced 50/50 distribution, you would likely attain far poorer results. This says nothing about your own predictive ability; it is the nature of learning on a skewed distribution in general.
Anecdotally, I do not know many women in software development, but I do know many men, most of whom I would consider very proficient at development. I have had the opportunity to review their code on many occasions and have generally found it well-documented and quite easy to follow (this is one of the reasons I consider them good developers, in fact). Since I don’t know many developers I would consider poor, this may be a result of my own selection bias, but I think we can agree that poor developers of either gender are generally undesirable for an organization in any case. This suggests that among the groups you really wish to have in the organization to begin with, men are just as capable as women of producing readable code.
To summarize, though I believe that more women in development would benefit the workplace due to the introduction of new perspectives into the organization, I do not believe that their code quality or readability is inherently different - nor do I believe that the assumption that it is constitutes a sufficient reason to introduce a gender bias into the selection process.
To conclude with a bit of irony, thank you for reading through my own male verbosity
Best Regards,
Michael Barnathan
June 16th, 2008 at 7:18 am
I think trying to spot the gender from some code excerpt is quite hard, usually written code shows more the mood of a person. I felt disgusted by the sexism in this broad generalization.
June 16th, 2008 at 8:13 am
Reposting (and rewording) what I wrote on Slashdot: The “showing off” is cross-gender and can work both ways.
Cross-gender — remember girls in high school? It wasn’t wheelies or backflips, but it was popularity and clothing, even who you were dating.
Working both ways: If you have a culture in which “cleverness” is desirable, then people will write clever (but unreadable) code. If you have a culture in which elegance is desirable, then people will write elegant, readable code — for exactly the same reasons.
My Slashdot comment: http://developers.slashdot.org/comments.pl?sid=584985&cid=23811601
June 16th, 2008 at 8:37 am
Sexism is boring.
June 16th, 2008 at 9:46 am
And how many deployments of Ingres are currently in the field…?
Your business would suffer less if you focused more on your department of “emerging technologies” instead of sensationalist horseshit. What have you done recently? An Eclipse DTP plug-in? Something, in a worst-case scenario, a solitary coder could do in 6 months, and get through QA in another month or so?
Seriously. Get with the program. Good coding practices are not based on a X or Y chromosome, and the plural of anecdote is not data. You should be ashamed of yourself.
June 16th, 2008 at 10:02 am
If a man said this about women it would be considered highly offensive and sexist and he’d burn in hell for writing it.
But I’m still going to say it, from what I’ve seen (which I do NOT claim to apply to all women, unlike the author does with men) women make terrible coders, they’re code can be “pretty” and “well-documented”, but most of the code I’ve seen was terribly long and inefficient, didn’t handle errors correctly and was eventually rewritten by men into short, efficient functions that handle errors perfectly.
And I’d much rather have good code than crap code with lots of explanation.
Also in your interview you say that the women mainly handle localisation and stuff, the men do most of the actual coding, of course there’ll be more documentation done by the woman, it’d take an idiot (or someone who doesn’t understand software development, like the audience of your horribly biased interview) to actually think that these figures hold up when comparing people doing completely different things.
I think you should never have said women and men, you should have said that programmers document less than those responsible for localisation, but that of course wouldn’t get your biased point across
June 16th, 2008 at 10:43 am
“women are more collaborative and, in a time of crisis, will pull a team together to get the problem resolved.”
Yeah, women are much more likely to have strong leadership skills and will sacrifice work/life balance to do anything it takes to resolve things. Women will never let a bad situation perpetuate, they are highly confrontational to adverse situations.
What planet are you on? These are blatantly male characteristics.
June 16th, 2008 at 10:58 am
[...] set off plenty of comments on that blog post and elsewhere, like this one at Slashdot and on McGratten’s own blog, where, by the way, she says her company’s code is “fairly androgynous” because [...]
June 16th, 2008 at 12:29 pm
You’ve got to be kidding me.
Is the meta-press you’re getting around this interview an intentional part of your “viral marketing” campaign, or are you really this dense?
The best part is where you say you can tell with 70%-80% accuracy whether the code you are looking at was written by a man or a woman at your company. But you also say that 80% of the coders are men.
Congratulations, you’re doing WORSE THAN CHANCE.
June 16th, 2008 at 8:26 pm
I actually did my preliminary masters work on just this subject. Evidence in sociolinguistics strongly supports the conclusion that men might write more obfuscated code. If you’re in for a bit of a slog
, you can take a look at the paper I wrote on the subject here:
http://www.uleth.ca/gendersymposium/Kim.doc
The Future Work section includes a lot of work that I would like to have done towards getting some more definitive findings together, but didn’t have the time/funding. If anyone’s interested in pursuing the research, I’d love to hear from you. I think this is an extremely valuable pursuit. Figuring out the actual differences between men’ and women’s code will go a long way towards breaking down the programming languages barriers.
June 17th, 2008 at 3:15 am
Emma,
I simply don’t believe that you hadn’t thought about this topic before.
I believe you’ve felt some of the heat resulting from your opinions and are trying to cover up.
If you didn’t have an opinion on gender related to work, you should have said so in the interview. Instead you gave very clear answers on how women are “better” and how you are trying to promote women into the workforce.
You are sexist; and should not be in a position of authority.
June 17th, 2008 at 3:32 am
Kim, any research that states such a biased viewpoint as fact and as a basis for interpretation of numbers; doesn’t deserve any time or money. A balanced view would have at least looked at equivalent factors in the male part of the equation. You didn’t even do that….
“The female half of the human species has customarily been discriminated against in most human societies [44]. Cultures around the world have assigned their females labels from houseslave [3] to demoness [39]. “
June 17th, 2008 at 6:49 am
@Kim Hansen- Here’s another way to break down the programming languages barriers: write your own language that will be suitable for women. It’s easy to do, just get a really smart woman programmer and have her create a language to your specifications. Let me know how that works out for you.
June 17th, 2008 at 7:38 am
In all due respect, that is the most sexist comments I have seen in a long time Emma. Good code is good code, and men as women can be as good (or bad) when it comes to code, engineering or architecture.
) is asexual, appreciated by all, and is truly a great equalizer. But then again it is a male opinion which you may disagree with.
Elegance is a state of mind and being French, Irish, or American makes no difference.
I truly wish there was more women in engineering and working a lot internationally, I must say that other culture such as India would strongly disagree with you. Code (like good Irish Whisky
Congratulation also on your success (by any scale)
June 17th, 2008 at 5:23 pm
I have to agree that these comments are extremely sexist and, yes, Paul, if a man had said them, the women would be screaming bloody murder about how men are so terrible to us…
There have been, are currently, and always will be differences between the way men and women think. However, there are also differences in personalities within the same sexes. Code is much like handwriting. One may be able to tell a gender difference with certain types, but, generally, it would be difficult, if not impossible, to tell male from female and vice versa. I am not a programmer or developer, I am a middle-aged woman with a degree in History, but I can figure out that saying something like this was as wrong as saying men are better than women at washing dishes or cooking.
Emma, honey, you can stroke your own ego without having to insult other people to do it. In my opinion, if you have to stroke your own ego, you’re not very good at what you do.