Open source: dangerous to computing education? (opensource.com)

Posted Feb 17, 2010 23:11 UTC (Wed) by farnz (subscriber, #17727)
In reply to: Open source: dangerous to computing education? (opensource.com) by shmget
Parent article: Open source: dangerous to computing education? (opensource.com)

LWN (unlike e-mail based projects such as the Linux kernel) actively encourages the use of nicknames, rather than full names. And "farnz" is closely related to my real name (Simon Farnsworth).

I've gone and looked at the names of posters in LKML threads, and names in the linux-2.6 git repository; almost all of them are using real names, many of which let me infer a likely gender, and in some cases, a likely cultural background. Now, I could be completely and utterly wrong in the assumptions I'm making (after all, they're built on my cultural background), but people do seem to use their real name when contributing to at least one major open source project.

I note also that you're attributing a hidden motive to my participation here - you previously said that I couldn't even think about guessing what gender, race, or religion the participants in open source projects might be, because it all takes place over the Internet. Now, you've gone to demanding that I show fault, and started to claim that I'm insisting that these problems be fixed by OSS; that was never my claim.

I am not demanding that projects change. I am asking them to be honest about their biases, which in some cases will require introspection; if a project genuinely wishes to participation by some groups, then it does not need to change at all. If it does want more women, or more Buddists, or more Zulus, and has a culture that deters them from participating, it needs to recognise that just because we're all communicating via computer doesn't imply that race/religion/gender issues are eliminated.

I also would like projects that are deterring some groups of potential contributors from getting involved to understand that they're doing so, and to ensure that the groups they're chasing away are deliberately selected. After all, not all groups of possible contributors are people you want to have involved - I'd rather not have the group "people who do not, and will never, understand C" working on the Linux kernel, for example.

Incidentally, your double-blind test study is not a good study; the point I'm making is that there is rarely (if ever) overt sexism or racism in successful open-source projects, and thus I would expect all projects to accept patches on merit; this doesn't mean that a project will attract or keep a diverse selection of contributors.

The problem is in the way the culture of projects interacts with the cultures of possible contributors, which results in some groups not participating at all; further, every project of significant size needs contributors to do more than fire patches at the project and ignore feedback. For example, when Al Viro deconstructs someone's abuse of the VFS interface, contributors who intend to submit things that use VFS APIs should be paying attention, and noting what he says. You therefore need your study to look at why some groups are underrepresented in open source as compared to closed source; from that study, you can determine what needs to change, and where the changes are needed.

Open source: dangerous to computing education? (opensource.com)

Posted Feb 18, 2010 0:19 UTC (Thu) by viro (subscriber, #7872) [Link] (1 responses)

Oh, lovely. Let me see if I got it right:
* we are assumed to have biases acting against mostly unspecified groups
* it's not the accusers' responsibility to provide any examples, let alone evidence.
* _we_ should come up with the list of such biases, or admit dishonesty.
* no, showing the lack of discrimination against a group is not enough (are you seriously saying that you can recognize Buddhists by email address, BTW?); the biases are supposed to be in culture being uncomfortable for group in question.

Frankly, it's getting SCO-worthy. The same kind of logics...

Open source: dangerous to computing education? (opensource.com)

Posted Feb 18, 2010 10:03 UTC (Thu) by farnz (subscriber, #17727) [Link]

I'm afraid you got it wrong:

Yes, you have biases acting against various groups. Some of them are clearly specified (e.g. Linus Torvalds is biased against people who think the kernel should be written in C++ not C). Others are unconscious, but you'll quickly spot them if you're looking for them. The worst are those that result from culture clashes.
No; the accuser should provide examples of groups they think you are biased against, with an explanation of why they think that it's you, not a general thing (e.g. 3% of open source contributors are women, 20% of closed source contributors are women).
Once you have been faced with an evidenced claim, you get two choices:
1. Try and identify your unconscious biases that lead to such a claim being true, and deal with them. This might be (for example), toning down your language when ripping apart a bad patch, or leaving a flame in a drafts folder for 24 hours, then rewriting it if you still want to send it.
2. Assert that you don't actually care; if the group that is underrepresented cares, it should do something about it, and if it wants specific changes from you, it has to ask for them and justify making you change.
Showing lack of discrimination is enough. However, that's not simply a case of showing that good patches will be accepted regardless of who submits them; you also need to show that people who write good code are getting involved regardless of their gender, religious beliefs, race or other irrelevant factors.

The point is that unreasonable discrimination is not just the obvious "ooh look. This patch is from A GURL! It's not worth reading." It's also subtler issues that put people off from contributing at all.

Open source: dangerous to computing education? (opensource.com)

Posted Feb 18, 2010 1:50 UTC (Thu) by shmget (guest, #58347) [Link] (13 responses)

"I could be completely and utterly wrong in the assumptions I'm making [...], but people do seem to use their real name when contributing to at least one major open source project."
All I can tell is that, in the linux kernel they use 'real looking' names. I have no way to know if 'Jim Smith' is really a man, as the firstname tend to indicate, or a 'nom de plume'.

"I note also that you're attributing a hidden motive to my participation here"
I'm sorry if that is the impression. that is not the intention.

"and started to claim that I'm insisting that these problems be fixed by OSS; that was never my claim."
Noted.

"...if a project genuinely wishes to participation by some groups, then it does not need to change at all."
False dichotomy: there is at least a third option: the 'I Don't care what the actual or perceived demographic mix is'

"If it does want more women, or more Buddists, or more Zulus"
Why would a project want that for the sake of it? Wouldn't a project want to attract the most talented persons it can, regardless of anything else ?

"it needs to recognise that just because we're all communicating via computer doesn't imply that race/religion/gender issues are eliminated."
Again, race/religion/gender is only visible insofar as the contributor want to make it visible.
I can participate in a OSS projet as Georges Brock, Daniel Sullivan, Yullila Tcheckovsky, Wei Pang Tsu, Dominique Durant, Hans Weiser, Amadou Ahidjo, etc... And you would still not know what my ethic background, my sex and even less my religion is... and quite frankly nor should you care.
(note: farnz could very well have stood for Farid Nariz, instead of Simon Farnsworth. As far as I'm concern that is completely irrelevant to the quality of your argument. Furthermore, for all I know you real name could be Maria Castello... How could I possibly know ? I this context, your name is what you say it is. Again, on the internet you are who you say you are, and it doesn't not have to be 'real')

"Incidentally, your double-blind test study is not a good study; the point I'm making is that there is rarely (if ever) overt sexism or racism in successful open-source projects, and thus I would expect all projects to accept patches on merit; this doesn't mean that a project will attract or keep a diverse selection of contributors."
If a project already accept contributions on merit, what else is there to do ? start to accept contributions based on other criteria than merit ?
(quite frankly, I would expect such a test study to show 'some' sexism and/or racism, certainly not 'overt' but still there. Participating in a OSS project does not magically shed everyone of their prejudices.)

Most of the objections as I understand them, are based on the discrimination, sexism of the potential contributor's cultural background (In my culture I'm expected to behave submissively(for example), therefore the OSS project must set the same expectation for me, or I won't fell comfortable - yes this is what "a culture that deters them from participating" means). Why should a tentative meritocracy goes out of its way to accommodate other's sexism/discrimination ?

"every project of significant size needs contributors to do more than fire patches at the project"
I used patch as a proxy, but 'contribution' is a more generic term that convey my intent better.

Open source: dangerous to computing education? (opensource.com)

Posted Feb 18, 2010 10:21 UTC (Thu) by farnz (subscriber, #17727) [Link] (1 responses)

We're, I think, agreeing violently. My point is that discrimination that results in an open source project not being a meritocracy can be subtle, and that projects should be aware of this, and should try and correct for their own subtle discrimination.

If I've understood you correctly, you're claiming that open source projects should be aiming to be true meritocracies, so that when the stats come up bad, it's a reflection not of open source culture, but of the wider world we find ourselves in.

These are not opposing viewpoints; unfortunately, some people in my camp have a nasty habit of assuming that until the stats are balanced, the project in which the stats are unbalanced has a problem. This is not my viewpoint; I argue that for as long as society as a whole has a problem, projects should watch themselves for signs that they've stopped being meritocracies.

Open source: dangerous to computing education? (opensource.com)

Posted Feb 18, 2010 23:17 UTC (Thu) by shmget (guest, #58347) [Link]

"projects should watch themselves for signs that they've stopped being meritocracies."

I do agree with that...

I just strongly disagree that the demographic breakdown of the contributor list is such a sign.
I also strongly disagree that the tone of a mailing list is any indication of the level of meritocracy achieved by a project.

Quite frankly, as far as I am concerned, the worse part about being called an idiot of mailing list is when I come to realize that I WAS indeed an idiot... otherwise, a Courteline quote comes to mine
"Passer pour un idiot aux yeux d'un imbecile est un delice de fin gourmet"

Open source: dangerous to computing education? (opensource.com)

Posted Feb 18, 2010 20:34 UTC (Thu) by DOT (subscriber, #58786) [Link] (10 responses)

I think you definitely want to attract women, just for the sake of it. They represent about 50% of the world population, and thus 50% of potential contributors. If you do something that deters women specifically, you handicap yourself. The same goes for any other demographic, but especially women are a huge resource that open source communities just miss out on.

Open source: dangerous to computing education? (opensource.com)

Posted Feb 18, 2010 20:42 UTC (Thu) by dlang (guest, #313) [Link] (1 responses)

there is a huge difference between working to avoid deterring anyone from contributing and saying that you want contributers of a certain flavor just for the sake of it.

Open source: dangerous to computing education? (opensource.com)

Posted Feb 18, 2010 22:01 UTC (Thu) by DOT (subscriber, #58786) [Link]

Theoretically maybe, but you need to know what you're doing wrong and turn that around by targeting those specific minorities. Just having some vague goal of allowing anyone to participate will not solve the very real problem.

Open source: dangerous to computing education? (opensource.com)

Posted Feb 18, 2010 23:01 UTC (Thu) by shmget (guest, #58347) [Link] (7 responses)

"They represent about 50% of the world population, and thus 50% of potential contributors."

Asian-American represent about 5% of the US population, African-America represent about 12% of the us population, therefore according to your argument, the NBA should have 12% f black and 5% of Asian, otherwise that indicate a 'very real problem'.
I would note also other groups that seems to have such 'very real problem', such as Country singers, Blues musician, Downhill Skier, Surfers (way under-represented in Swiss, and other inhabitant of land-locked countries), Chess players, etc...

Theses kind of fallacious use of stats are really annoying.

Open source: dangerous to computing education? (opensource.com)

Posted Feb 18, 2010 23:58 UTC (Thu) by shmget (guest, #58347) [Link] (3 responses)

Oh I forgot another 'organization' with such a 'very real problem': Health-care providers.
In the US, among the 90 years old or older, there is only 24% of men.
Surely that indicated that Health-care providers are sexist and that there is a 'very real problem' that need to be addressed their indisputably proven bias against men.

Open source: dangerous to computing education? (opensource.com)

Posted Feb 19, 2010 0:54 UTC (Fri) by njs (guest, #40338) [Link] (2 responses)

What? No-one thinks that it is always, inherently, a problem to have a gender imbalance in every domain. (This just in! Very few men bear children! Also, very few women are diagnosed with prostate cancer! Oh noes!) That's a straw-man you just made up.

A gender imbalance is a clue -- a symptom -- that shows there must be something going on to create the imbalance -- especially when the imbalance is as ridiculously huge as it is in FOSS. The question then is whether that "something" is something we want to support. In the case of health care providers, well, women tend to live a bit longer, doesn't seem to have anything to do with the health care providers. In the case of FOSS, well, our communities contain many people who drive away women by being jerks to them, and many more people who stand by and let it happen. (Seriously, this is *copiously* documented, see any of the formal studies, talk to any woman in FOSS, read any previous LWN article/discussion on the subject, etc.) That's a little different.

For instance, there are people who -- whenever this issue comes up -- start flaming, making up straw men, and otherwise trying to derail reasoned discussion. This kind of nonsense is very effective at telling potential developers that any problems they have will be ignored or mocked, and that their presence is not valued, so they go somewhere else.

Open source: dangerous to computing education? (opensource.com)

Posted Feb 19, 2010 4:33 UTC (Fri) by shmget (guest, #58347) [Link] (1 responses)

"That's a straw-man you just made up."
"doesn't seem to have anything to do with the health care providers"
That was not a straw-man, that was sarcasm...
It didn't realize that I needed <sarcasm/> tag around that.

A gender imbalance is a [..] a symptom...
aka a weasel word. Labeling it a 'symptom' is a Begging the question fallacy. You postulate your conclusion in in the premise.

"In the case of FOSS, well, our communities contain many people who drive away women by being jerks to them"
Nope. There are certainly people who are jerk, but they are mostly equal-opportunity jerk. Since everyone is subjected to the same amount of 'jerkitude', it is not discriminatory.
Claiming that women are poor soft thing that can take it, in such a drastic fashion that it explain the demographic ratio, is actually insulting to women.

Again, referring to previous post. Establish that contributions in a project are rejected on the ground of sex/race or other irrelevant factor.. then you'll have a case that that project need adjusting.

"(Seriously, this is *copiously* documented, see any of the formal studies,"
I've read them. (well, I've read the one done under the auspice of the EC, a couple of years back, IIRC)
Most 'arguments' in that study concerned society at large and had nothing to to with FOSS. The numbers were seriously twisted, and many time the conclusions of the author(s) were worded DESPITE what the numbers actually showed.
(in other word: make an hypothesis, test it and when the test failed to prove the hypothesis, just ignore it and claim that the hypothesis still stand.)

"talk to any woman in FOSS" And how would I know that I'm 'talking' to a woman, are you one ? Should I take your word for it ? Am I one ? Please indulge me with your expert guess.
I would find it extremely weird to actually inquire on the sex/age/ethnicity/religion/... of a poster in some Dev related mailing-list. How could that kind of information be relevant ?

"For instance, there are people who[...]
Nice passive-aggressive tirade.

The whole "watch you language, there are women around" lobbying is highly condescending. I find that kind of Dark-aged chivalry much more offensive than the occasional sexist jerk. Women do not need a FOSS paralympic to win medals.

Open source: dangerous to computing education? (opensource.com)

Posted Feb 19, 2010 5:29 UTC (Fri) by njs (guest, #40338) [Link]

> Nope. There are certainly people who are jerk, but they are mostly equal-opportunity jerk. Since everyone is subjected to the same amount of 'jerkitude', it is not discriminatory.

This is not true. Again, ask the women who participate. The EC study you dislike found that 80% of women who responded had noticed sexist behavior. I.e., non-equal-opportunity jerkitude. Or, y'know, look at any of these threads.

> Claiming that women are poor soft thing that can take it, in such a drastic fashion that it explain the demographic ratio, is actually insulting to women.

Nonsense. I couldn't take half the stuff some of the women who participate have to put up with (which goes up to and including death threats). I'm not still here because I have a tougher skin, I'm still here because I don't need one.

> "talk to any woman in FOSS" And how would I know that I'm 'talking' to a woman, are you one ? Should I take your word for it ? Am I one ? Please indulge me with your expert guess.

So, uh, how is that a response? Is the claim that since you can't find any women (and obviously you've looked very hard, it's not like they have blogs and give keynotes on sexism in FOSS and maintain wikis documenting sexist behavior or anything), then what they say is irrelevant?

> The whole "watch you language, there are women around" lobbying is highly condescending. I find that kind of Dark-aged chivalry much more offensive than the occasional sexist jerk.

You're right, that's a pretty sexist way of looking at things. But you're the one who's reading that into my comments. I'm just saying that acting like sexist jerks causes harm to people and communities. "Try not to do harm" is not a particularly dark-aged value, AFAIK. YMMV.

And did you really just claim the right to judge some sorts of sexism particularly offensive, and others relatively harmless, on behalf of those women you can't find to talk to? That's very, well... chivalrous of you.

Open source: dangerous to computing education? (opensource.com)

Posted Feb 19, 2010 19:36 UTC (Fri) by DOT (subscriber, #58786) [Link] (2 responses)

So, are you saying that women have a natural tendency to not participate in FOSS? I can't believe that's what you're saying. What do you think accounts for the low numbers of women in FOSS?

What I'm saying, is that women and men have an equal tendency to get interested in FOSS. So following from that, there should be around 50% women in FOSS. The *fact* that this is NOT the case means we're missing out on that HUGE resource.

That's a huge problem, because we would almost double our number of contributors if we could attract more women.

Open source: dangerous to computing education? (opensource.com)

Posted Feb 20, 2010 0:40 UTC (Sat) by shmget (guest, #58347) [Link] (1 responses)

"What I'm saying, is that women and men have an equal tendency to get interested in FOSS."
I understand that is what you're saying. But do you have any data to back-up that claim, or is it just wishful thinking ?

"So following from that, there should be around 50% women in FOSS."

Yet it is not the case, so _maybe_ your postulate is wrong ?

Let me try another way:

s/FOSS/Chess/ in your statement... what do you conclude ?
(note: there is 1 woman in the top 100 chess players as of January 2010. I don't know why, but that is the case.)

"because we would almost double our number of contributors if we could attract more women."
99% (1) of the men (and 99.9% of the women) do not participate in FOSS, Why do you think it is more pertinent to concentrate of the 0.9% gap, rather than the 99% un-taped resource.

(1): I could not found data. I picked a number. 67 millions distinct OSS contributors worldwide sound like a generous estimate. But the same argument would hold with 90% - 99% - and I quite certain that there is not 670 millions of open source contributors.

Open source: dangerous to computing education? (opensource.com)

Posted Feb 20, 2010 14:34 UTC (Sat) by DOT (subscriber, #58786) [Link]

You can try it any way you want, but you're avoiding the issue. Women are not somehow less interested in FOSS. Asians are not somehow less interested in FOSS. Muslims are not somehow less interested in FOSS. They are very normal people with the same kind of interests as Christian white males.

What they are, is scared away by a crowd of otherwise very normal people who just don't know that they make minorities uncomfortable. And the fact is, if you are aware of that issue, you can double the community. Tell me again how you are going to double the FOSS community.

I don't care about chess, so I don't care to figure out if there is a problem there.