WTF Is Wrong with Open Source Communities?
blog.container-solutions.com
blog.container-solutions.com
To whom? You can always start your own FOSS project and include more varied people/genders. You cannot dictate what others do on their own time/dime as a passion project/hobby.
Did these people attempted to get into OSS and failed, were not welcomed, etc?
That would be a criterion that makes sense. Else it's either a non-issue (they simply elect not to, for whatever cultural or other reasons) or an upstream issue (e.g. educational).
People feeling not welcome in some FOSS project can always start their own project (and be as inclusive as they wish).
If Lynne Jolitz could get into FOSS in 1992, others can do it in the much more sensitive and inclusive tech/general culture of 2021.
... or why ?
1) Diversity of people, of experiences and point of views can boost a project (this includes "minorities", genders, countries of origin... but it is not limited to this). 2) Being able to attract more people to contribute can boost a project.
So if you care about an OSS project or OSS in general, that makes sense to think about those issues.
On a bigger picture, people working (often for free) on OSS are doing an amazing job that has a big impact in the world we are living. In a more societal point of view, it also make sense to 1) try to include the views and experience of many different people on project that will have an impact on everybody 2) find ways that more people and more diverse people contribute
Everybody can contribute on an OSS. But there are things preventing some people to contribute and that we can ameliorate. There are ways to make more people and more diverse people to contribute or to contribute more.
Maybe, but that's an abstract idea about society/culture, not some natural law that applies to everything.
The maintainers of a project don't have to necessarily agree, or want to dillute their own perspectives. FOSS is often about personal (or small team) passion and/or a BDFL driving it, not about "design by commitee" (and the commitee being diverse can make things worse). Many FOSS projects want or could use less "points of view" and more singular vision and focus.
Open source contributors self-select. They have to put time in, usually for $0 and often without thanks. I don't know of any FOSS project who asks, or cares, about race or sex or anything else except code, or if the project is ill-maintained, at least your patch will get just as ignored regardless of your background.
The article reads like some (white) Karen who wants to talk to FOSS' manager, there isn't any.
If you think there should be more people sending patches or running FOSS projects, bring them, if their contributions are any good non-broken projects will be glad for them.
They have to put time in, usually for $0 and often without thanks > Right. As the author states in a comment below the piece: "Open source work should be paid work, and, until it is, it shouldn't be a requirement to getting a job -- that only increases inequity when so much of the unpaid care work falls on those already minoritized by tech and open source."
I don't know of any FOSS project who asks, or cares, about race or sex or anything else except code, or if the project is ill-maintained, at least your patch will get just as ignored regardless of your background. > So how do you explain all the research that tells you this isn't true? Such as this piece - https://peerj.com/preprints/1733/ - cited in the article.
Also the users the data observes are found on Github and Google+. This is a insufficient data if you write about the open source community in general.
The implicit bias test the authors used to certify themselves have problems justifying their relevance.
Ie, not in the article.
Can you elaborate on how that paper 'tells me' my own lived experience the last 20 years maintaining FOSS projects and offering patches 'isn't true'? That sounds arrogant on your part.
The paper tries to guess gender from github nyms and profile pics, and looking for google + integration, this seems like a red flag to me. I couldn't guess gender or sex from that, especially considering regional differences, like anime pictures for profile pics, unless it was explicitly telling me in the name. What they may instead be partitioning is the aggressiveness / confidence of the nym and its presentation, which corrolates later to what happens at the first hurdle for their patches.
For simple bug fixes, patches are typically acceptable right away. For more complex fixes and new features, some negotiation is needed with the maintainer as to how it should be done, there are often several ways that are reasonable, but eg, only one or two that align with future plans for the code. If the maintainer asks for a redo, a good proportion of contributors give up, they have exhausted their initial energy. Similarly, when asked for tests, or presented with considerations for other supported platforms the contributor doesn't care about, more give up. It requires extra confidence to approach the maintainer about preferred ways to do the patch before just starting on it. The maintainers are sometimes assholes (I am pretty sure, to everyone).
I think it's very hard to unpick why patches didn't go anywhere unless you follow the story of each, certainly you need abnormal energy and confidence to get started and get over the hurdles for anything nontrivial.
You are looking only at the last step into becoming a FOSS contributor. There are other selection filters you have missed, but which are touched on in the essay.
The article mentions "the vast majority of open source contributors are either volunteers or work for sponsoring organizations." I believe that is correct.
Those who work for sponsoring organizations aren't quite "self-selected", in that their employer is also involved, if only to grant copyright permission. I wouldn't say that all of the Google staff working on the open-source part of Chrome are self-selected open source contributors. Would you?
And don't forget the article mentions a Fortune analysis that "the world’s biggest companies, who are the frequent sponsors of high-profile OSS projects, are still lagging behind in diverse hires." That's a selection filter right there.
Even if all FOSS contributors are self-selected, there are clearly systemic biases in who is able to make a decision.
Consider the mentioned "unpaid care work", which includes raising children and taking care of the elderly. Many societies regard this as a woman's responsibility, so even if two people have to decide to spend their volunteer time working on open source vs. doing unpaid care work, a man in those cultures may find it easier to "self-select" open source development than a woman might.
> If you think there should be more people sending patches or running FOSS projects, bring them
Someone working two jobs to support a family has little free time for open source development. Someone working part-time at a job and full-time child minding also has little free time. OTOH, someone with money can have a job with the evenings off, can pay for a babysitter, can order food instead of cooking, etc.
As a result, poor people are underrepresented in open source software development if only because of the lack of available free time.
Historic racism in the US contributes to the modern correlation between wealth and the social construct of "race". Blacks, for example, are more likely to be poor, and therefore less likely to be in a position to be able to "self-select" in the first place.
Personally, I see Universal Basic Income as a way to get a lot of good open source developers, by increasing the number of people who are able to volunteer.
> I don't know of any FOSS project who asks, or cares, about race or sex or anything else except code
Have you missed the last 10 years of discussions about FOSS projects that have actively removed the use of "master/slave"? That sure sounds like FOSS projects which (at least claim to) care about race.
Here too you are looking only the end point of the process - someone who contributes to a FOSS project. Those who have removed "master/slave" argue that such terms may inhibit people from becoming a FOSS contributor in the first place.
Similarly, a FOSS project with a Discord project discussion channel where the developers regular post pictures of sexy women might disproportionately inhibit contributions from women.
Even if they are completely wrong, my point is that there are many selection filters beyond "does a project accept or reject code based on the contributors race, sex, etc."
Yeah, so because of that they don't 'select' themselves to contribute to FOSS. If they don't have a PC, or can't program, likewise. I'm not responsible for any of that, as a FOSS maintainer, they turn up with patches in their hand, or they don't. If they don't, it's totally out of scope for me. With FOSS I don't even have a way to know such users exist.
If they want to contribute something, I can spend my time assisting that to happen, or not, that's the only button I can push.
>> cares > slave
I mean cares about race in the context of treating their patch differently because of it, this is the implication of the article. Maintainers are motivated to take in useful code.
> developers regular post pictures of sexy
If devs want to do that, they're not breaking any law. But a lot of contributors (eg, ones logging in on a work pc) would find it intolerable to the project's detriment. They'd lose me at tne word 'discord' anyway.
> If they don't, it's totally out of scope for me
This article about "Open Source Communities" and "Open source leadership".
There's plenty of FOSS projects which have no communities. I have several. I don't need a CoC since I'm the only point of contact.
Do you want a project community? I don't. I'm a lousy community manager.
If you do, which people do you want to encourage to be in that community? What social practices do you want to have frowned upon? If you have no limits, what do you do about griefers? If people break those limits, what do you do?
> I mean cares about race in the context of treating their patch differently
I understand. But why not also care about people who could submit patches but don't?
> intolerable to the project's detriment
Good point! Losing possible corporate contributions due to NSFW images is very similar to, say, losing possible voluntary contributions for having a rude and aggressive management style, or for other harassing community practices.
I'll try to remember that example in future discussions - thanks!
If your open source project is for business, you can of course chose to have any content restriction you want. But there cannot be a general rule. Projects on github are endangered enough to be both, abused and helped by corporate input. I don't the MS embrace care program to be extinguished in phase 2.
Most projects are projects of passion and if that passion includes pictures of naked people, who am I to argue against that?
I'm not arguing for or against any type of project. Even those "mostly due to other reasons" include Bona Fide Occupational Qualifications; a FOSS project for nudist organizations may include pictures of naked people and be developed at someone's work even within EEOC guidelines.
What I am arguing against are the ideas that, when talking about "WTF is Wrong with Open Source Communities": 1) "self-selection" is all that meaningful, esp. as the piece mentions issues with that concept, and 2) lack of discrimination of patch acceptance due to submitter's protected group membership status is the primary standard.
I do think there are factors, large ones even, that keep some women out of tech and they aren't sourced in biology. None of these are adressed by the suggestions in the article.
I think the largest expression of sexism actually comes from people that suggest "solutions" like this. Using stereotypes now to make it clear in how I see positions like these. They wouldn't ever go into a nail saloon, complain about legal security or their behavior and cite it as the reason why they have difficulties hiring males. But it is with males that you insinuate discrimination. That sounds very well like prejudice to me and I also think the referenced studies came to prederterminded conclusion and are of little scientific value.
Most open source projects don't contain nudes, but some actually do. I don't think this is a problem. To think this is the reason women are kept out of the industry is nonsensical to me. I am missing even the hint of truth here.
Source please? Last I heard of, that was not the case.
As https://www.wired.com/2017/06/diversity-open-source-even-wor... contexualizes about the research published at https://opensourcesurvey.org/2017/ :
] GitHub, the world's leading repository of open-source code, surveyed 5,500 open source users and developers from around the world on a range of topics. It also asked for demographic information. And it was informative. Of that randomly selected cohort, a full 95 percent of respondents were male. Only three percent identified as female and one percent as non-binary. According to Bureau of Labor Statistics, about 22.6 percent of professional computer programmers are female. About 16 percent of respondents said they belonged to ethnic or national groups that are in the minority in the country they live in. Black, Asian, and Latino programmers account for a total of about 34 percent of programmers in the US, according to the bureau.
This is one of the studies referenced in the paper.
To be fair, these numbers (world-wide participation vs. US labor representation) are not directly comparable. But since a large number of FOSS developers come from the US, there would need to be a huge number of non-US male software developers to shift the demographics meaningfully.
Other counter-examples:
> Several surveys have found that existing FOSS communities are very homogenous populations and made up of mostly men. There are significantly fewer women participating in FOSS when compared to the percentage of women in computing in general. - https://ir.library.oregonstate.edu/concern/graduate_thesis_o...
> Although organizations are taking actions to increase gender diversity, the percentage of women in OSS projects are in average lower than 10%. Only 7.5% of the contributions to public code from the last 50 years were authored by women [6]. Women represent only 5.2% of the contributors to the Apache Software Foundation [7], 9.9% in Linux kernel [8],and 10% of OpenStack contributors [9], three of the largest and most well-known OSS communities. Indeed, women represent only 9% of GitHub users [10]. - https://arxiv.org/pdf/2103.08763.pdf
and here's the US statistics, from https://www.bls.gov/cps/cpsaat11.htm :
Computer programmers: 417k employed, 21.1% women
Software developers: 1,883k employed, 19.4% women
Software quality assurance analysts and testers: 82K employed, 25.1% women
And you're the one who brought up "naked people". The point wasn't that NSFW images keep women out of FOSS, it was an example of how FOSS project participation CANNOT be viewed as "self-selection" because, as fivelessminutes correctly points out "a lot of contributors (eg, ones logging in on a work pc)" would be in workplace environments which prohibit certain types of imagery, like "sexy women", which was the term I used.
To turn that into "the reason women are kept out of the industry" is completely misreading my comments.
Women aren't discriminated against joining CS, they just decide themselves against it. The discrimination explanation does not hold up under scrutiny. But furthermore, it is quite easy to explain the OSS numbers on Github. It is pretty much exactly mirrors in the CS pipeline.
You are also a bad patron for those few women that are in OSS. Because the demanded formalities aren't meant to reduce discrimination, it is about big tech "embracing" OSS in my opinion. That is based on observations vastly nearer to reality than the discrimination story. So, wtf is wrong with the author?
And even IF numbers would be different between pipeline and OSS participation, it doesn't allow for the conclusion that were drawn. They are made up in my opinion.
You described your country's employment statistics, although without evidence.
You haven't listed the breakdown of gender representation in your country's OSS participation, so there's nothing to compare.
And "Europe" is rather diverse region.
As counter-evidence, I point you to https://blog.honeypot.io/women-in-tech-germany/ which claims women make up 16.58% of IT in Germany, and that percentage is lower than 15 other European countries.
To be sure, I can't vouch for the quality of that research, but so far you have presented none.
> they just decide themselves against it
Again, where is your evidence. I have already in this thread listed reasons for why women (who have more pressure to be unpaid care givers) and the working poor (which in the US is correlated with minority status) may have LESS choice than average, so saying "just decide themselves" ignores the factors which go into that decision.
> "It is pretty much exactly mirrors in the CS pipeline."
Evidence please. Since the citations I gave show otherwise.
> it doesn't allow for the conclusion that were drawn
Which specific conclusions do you disagree with and what evidence do you have besides your personal observations and biases?
> women and other gender minorities
Women aren't a gender minority.
> Enforced codes of conduct matter
I don't think so, you want OSS to cater to you. You sell yourself to companies that want to enforce compliance and you are the useful vehicle. This is why you get animosity, not because of your sex.
> has found that women strongly consider open source codes of conduct before making their first contributions
I dismiss this because there is no evidence at all.
> Open source leadership
You didn't even understand the concept.
> Zainab Daodu says you have to make an example of them.
I think I need to make an example of people trying to enforce some compliance or demands without any contribution, a common problem in OSS.