Why I won’t be attending Systems We Love

Systems We Love is a one day event in San Francisco to talk excitedly about systems computing. When I first heard about it, I was thrilled! I love systems so much that I moved from New Mexico to the Bay Area when I was 23 years old purely so that I could talk to more people about them. I’m the author of the Kernel Hacker’s Bookshelf series, in which I enthusiastically described operating systems research papers I loved in the hopes that systems programmers would implement them. The program committee of Systems We Love includes many people I respect and enjoy being around. And the event is so close to me that I could walk to it.

So why I am not going to Systems We Love? Why am I warning my friends to think twice before attending? And why am I writing a blog post warning other people about attending Systems We Love?

The answer is that I am afraid that Bryan Cantrill, the lead organizer of Systems We Love, will say cruel and humiliating things to people who attend. Here’s why I’m worried about that.

I worked with Bryan in the Solaris operating systems group at Sun from 2002 to 2004. We didn’t work on the same projects, but I often talked to him at the weekly Monday night Solaris kernel dinner at Osteria in Palo Alto, participated in the same mailing lists as him, and stopped by his office to ask him questions every week or two. Even 14 years ago, Bryan was one of the best systems programmers, writers, and speakers I have ever met. I admired him and learned a lot from him. At the same time, I was relieved when I left Sun because I knew I’d never have to work with Bryan again.

Here’s one way to put it: to me, Bryan Cantrill is the opposite of another person I admire in operating systems (whom I will leave unnamed). This person makes me feel excited and welcome and safe to talk about and explore operating systems. I’ve never seen them shame or insult or put down anyone. They enthusiastically and openly talk about learning new systems concepts, even when other people think they should already know them. By doing this, they show others that it’s safe to admit that they don’t know something, which is the first step to learning new things. They are helping create the kind of culture I want in systems programming – the kind of culture promoted by Papers We Love, which Bryan cites as the inspiration for Systems We Love.

By contrast, when I’m talking to Bryan I feel afraid, cautious, and fearful. Over the years I worked with Bryan, I watched him shame and insult hundreds of people, in public and in private, over email and in person, in papers and talks. Bryan is no Linus Torvalds – Bryan’s insults are usually subtle, insinuating, and beautifully phrased, whereas Linus’ insults tend towards the crude and direct. Even as you are blushing in shame from what Bryan just said about you, you are also admiring his vocabulary, cadence, and command of classical allusion. When I talked to Bryan about any topic, I felt like I was engaging in combat with a much stronger foe who only wanted to win, not help me learn. I always had the nagging fear that I probably wouldn’t even know how cleverly he had insulted me until hours later. I’m sure other people had more positive experiences with Bryan, but my experience matches that of many others. In summary, Bryan is supporting the status quo of the existing culture of systems programming, which is a culture of combat, humiliation, and domination.

People admire and sometimes hero-worship Bryan because he’s a brilliant technologist, an excellent communicator, and a consummate entertainer. But all that brilliance, sparkle, and wit are often used in the service of mocking and humiliating other people. We often laugh and are entertained by what Bryan says, but most of the time we are laughing at another person, or at a person by proxy through their work. I think we rationalize taking part in this kind of cruelty by saying that the target “deserves” it because they made a short-sighted design decision, or wrote buggy code, or accidentally made themselves appear ridiculous. I argue that no one deserves to be humiliated or laughed at for making an honest mistake, or learning in public, or doing the best they could with the resources they had. And if that means that people like Bryan have to learn how to be entertaining without humiliating people, I’m totally fine with that.

I stopped working with Bryan in 2004, which was 12 years ago. It’s fair to wonder if Bryan has had a change of heart since then. As far as I can tell, the answer is no. I remember speaking to Bryan in 2010 and 2011 and it was déjà vu all over again. The first time, I had just co-founded a non-profit for women in open technology and culture, and I was astonished when Bryan delivered a monologue to me on the “right” way to get more women involved in computing. The second time I was trying to catch up with a colleague I hadn’t seen in a while and Bryan was invited along. Bryan dominated the conversation and the two of us the entire evening, despite my best efforts. I tried one more time about a month ago: I sent Bryan a private message on Twitter telling him honestly and truthfully what my experience of working with him was like, and asking if he’d had a change of heart since then. His reply: “I don’t know what you’re referring to, and I don’t feel my position on this has meaningfully changed — though I am certainly older and wiser.” Then he told me to google something he’d written about women in computing.

But you don’t have to trust my word on what Bryan is like today. The blog post Bryan wrote announcing Systems We Love sounds exactly like the Bryan I knew: erudite, witty, self-praising, and full of elegant insults directed at a broad swathe of people. He gaily recounts the time he gave a highly critical keynote speech at USENIX, bashfully links to a video praising him at a Papers We Love event, elegantly puts down most of the existing operating systems research community, and does it all while using the words “ancillary,” “verve,” and “quadrennial.” Once you know the underlying structure – a layer cake of vituperation and braggadocio, frosted with eloquence – you can see the same pattern in most of his writing and talks.

So when I heard about Systems We Love, my first thought was, “Maybe I can go but just avoid talking to Bryan and leave the room when he is speaking.” Then I thought, “I should warn my friends who are going.” Then I realized that my friends are relatively confident and successful in this field, but the people I should be worried about are the ones just getting started. Based on the reputation of Papers We Love and the members of the Systems We Love program committee, they probably fully expect to be treated respectfully and kindly. I’m old and scarred and know what to expect when Bryan talks, and my stomach roils at the thought of attending this event. How much worse would it be for someone new and open and totally unprepared?

Bryan is a better programmer than I am. Bryan is a better systems architect than I am. Bryan is a better writer and speaker than I am. The one area I feel confident that I know more about than Bryan is increasing diversity in computing. And I am certain that the environment that Bryan creates and fosters is more likely to discourage and drive off women of all races, people of color, queer and trans folks, and other people from underrepresented groups. We’re already standing closer to the exit; for many of us, it doesn’t take much to make us slip quietly out the door and never return.

I’m guessing that Bryan will respond to me saying that he humiliates, dominates, and insults people by trying to humiliate, dominate, and insult me. I’m not sure if he’ll criticize my programming ability, my taste in operating systems, or my work on increasing diversity in tech. Maybe he’ll criticize me for humiliating, dominating, and insulting people myself – and I’ll admit, I did my fair share of that when I was trying to emulate leaders in my field such as Bryan Cantrill and Linus Torvalds. It’s gone now, but for years there was a quote from me on a friend’s web site, something like: “I’m an elitist jerk, I fit right in at Sun.” It took me years to detox and unlearn those habits and I hope I’m a kinder, more considerate person now.

Even if Bryan doesn’t attack me, people who like the current unpleasant culture of systems programming will. I thought long and hard about the friendships, business opportunities, and social capital I would lose over this blog post. I thought about getting harassed and threatened on social media. I thought about a week of cringing whenever I check my email. Then I thought about the people who might attend Systems We Love: young folks, new developers, a trans woman at her first computing event since coming out – people who are looking for a friendly and supportive place to talk about systems at the beginning of their careers. I thought about them being deeply hurt and possibly discouraged for life from a field that gave me so much joy.

Come at me, Bryan.

Note: comments are now closed on this post. You can read and possibly comment on the follow-up post, When is naming abuse itself abusive?

Ban boring mike-based Q&A sessions and use index cards instead

If you’ve ever been to a conference, you know the problem: A brilliant and engaging talk is coming to a close, and already a line of fanatic wild-eyed people (okay, mostly men) is forming at the audience microphone. Just by looking at them you know they will inevitably start their questions with, “This is more of a comment than a question, but…” Actually, you are grateful for the ones who are that self-aware, because most of them seem to genuinely believe that their barely disguised dominance play or naked self-promotion is an actual question that the rest of the audience would like to hear the answer to. So you scooch down lower in your seat and open your Twitter client so you can complain about how awful Q&A sessions inevitably are.

Fortunately, there is a way to prevent this situation entirely! Here is the formula:

  1. Throw away the audience microphones.
  2. Buy a pack of index cards.
  3. Hand out the cards to the audience before or during your talk.
  4. Ask people to write their questions on the cards and pass them to the end of the row.
  5. Collect the cards at the end of the talk.
  6. Flip through the cards and answer only good (or funny) questions.
  7. Optional: have an accomplice collect and screen the questions for you during the talk.

Better yet, if you are a conference organizer, buy enough index cards for every one of your talks and tell your speakers and volunteers to use them.

Why is the typical line-at-the-mike style of audience question so productive of bad questions? To start with, it gives the advantage to people who aren’t afraid to put themselves forward first and rush to the mike first. This means most or all of the questions are from people with relatively little self-doubt and a high opinion of themselves. Another draw for the self-centered overconfident type is the chance to be the center of attention while asking the question using the audience microphone. Then there is the lack of built-in limit on the time the purported question-asker is speaking. Finally, there is no way to screen the question for quality until the question has been fully asked (sometimes taking minutes). The end result is a system that practically invites self-centered, overconfident, boring, long-winded people to dominate it. (And you wonder why women almost never ask questions at your conference?)

By contrast, writing questions on index cards appeals more to quiet, thoughtful, self-effacing folks who are considerate of those around them. It allows you to screen the questions for quality. It limits the length of the question. It encourages actual genuine requests for clarification on the subject of your talk.

Get rid of line-at-the-mike style Q&A sessions. Replace them with index cards. Your conference attendees will thank you.

A personal appeal for support from Valerie Aurora, executive director of the Ada Initiative

This post originally appeared on the Geek Feminism blog and is reposted here for posterity.

Valerie Aurora

I’m writing to ask you to donate to the Ada Initiative.

A year ago, a friend of mine was groped at an open source conference. Again. I’ve personally been groped twice at conferences myself.

But what shocked me most was the reaction to her blog post about it. Hundreds of people made comments like, “Women should expect to get groped at conferences,” and “It was her fault.” Many of these people were members of the open source community. Some were even prominent leaders – that I was forced to work with directly in my job as a Linux kernel developer! I realized I’d felt alienated, unwelcome, and unsafe as a woman in open source for many years. I was furious and determined to make a difference.

So I quit my job and co-founded the Ada Initiative with Mary Gardiner. We are the only non-profit dedicated solely to increasing the participation of women in open source, Wikipedia, fan culture, and other areas of open technology and culture. Currently, women make up only 2% of the open source community, and 9% of Wikipedia editors, down from 13% a year ago. We want to change these trends.

You can help by donating or by spreading the word about our donation drive now:

Donate now!

Help spread the word

We’re proud of what we’ve accomplished already. Since our founding in early 2011, we helped over 30 conferences and organizations adopt an anti-harassment policy, organized the first AdaCamp unconference, provided free consulting on high-profile sexist incidents, wrote and taught two workshops on supporting women in open tech/culture, and ran two surveys, among other things.


We need your help to achieve our upcoming goals. The Ada Initiative is funded entirely by donations. Without your financial support, the Ada Initiative will have to shut down in early 2012.


Your donations will fund upcoming projects like: Ada’s Advice, a comprehensive guide to resources for helping women in open tech/culture, Ada’s Careers, a career development community, and First Patch Week, where we help women create and submit their first open source patch. You can learn more about how the Ada Initiative is organized and operated on our web site and blog.

Whether or not you can donate yourself, you can help us by spreading the word about our fundraising drive. Please tell your friends about our important work. Email, blog, add our donation button to your web site, and tweet. You don’t have to stand on the sidelines any longer. You can help women in open technology and culture, starting today.

Example conference anti-harassment policy turns one year old

November 29 was the one year anniversary of the publication of the example conference anti-harassment policy! One year later, over 30 conferences have adopted an anti-harassment policy of one kind or another.

This post originally appeared on the Geek Feminism blog and is reposted here for posterity.

This is a cross-post from the Ada Initiative’s blog. The Ada Initiative has put a lot of effort into helping conferences understand and adopt some form of anti-harassment policy. Your donations will help us continue to promote the policy and do similar work. Thanks!

November 29 was the one year anniversary of the publication of the example conference anti-harassment policy! Inspired by multiple reports of groping, sexual assault, and pornography at open tech/culture conferences, the Ada Initiative co-founders helped write and publish an example conference anti-harassment policy for modification and reuse by conference organizers. This example was the collaborative effort of many different conference organizers and community members, who all deserve thanks and credit.

One year later, over 30 conferences have adopted an anti-harassment policy of one kind or another. “More than 30” is a rough lower bound; several organizations have adopted a policy for all their events and run a dozen or more events per year. Some of the organizations that have announced that all their conferences will have a policy include Linux Foundation, ACM SIGPLAN, and O’Reilly (pledged). Here’s why some conferences have adopted a policy, in the organizers’ own words:

Tim O’Reilly: “[…] It’s become clear that this is a real, long-standing issue in the technical community. And we do know this: we don’t condone harassment or offensive behavior, at our conferences or anywhere. It’s counter to our company values. More importantly, it’s counter to our values as human beings.”

Jacob Kaplan-Moss, co-organizer of PyCon US, speaking for himself in this post: “A published code of conduct tells me that the conference staff cares about these issues, takes them seriously, and is waiting and willing to listen if an incident happens. It’s by no means a solution to the depressing homogeneity of technical communities, but it’s a step in the right direction.”

ACM SIGPLAN: “This policy has been in the works at the ACM SIGPLAN for several months; SPLASH 2011 is proud to be both the driver for that effort and the first ACM conference with such policy in place. This policy is not a symbolic gesture, delivered to satisfy a perceived need for political correctness, but instead goes to the core of both our personal beliefs and the beliefs of the community as a whole.”

Like any good open source project, the policy has been forked, adapted, and rewritten from scratch several times. Conference organizers looking to adopt a policy now can choose from several different policies. Many policies are linked to from this list of conferences with a policy; if you know a conference that is missing, please add it!

Some history

Why write an example anti-harassment policy? What we discovered after a little research (aided by the timeline of sexist incidents in geek communities) was the following:

  • Often, the person doing the groping, harassing, or showing of pornography honestly believed that their behavior was acceptable for the venue. Just as often, many other people went on record agreeing with them.
  • People who saw these incidents didn’t know how to respond to these incidents or weren’t sure who to report them to.
  • Conference organizers sometimes didn’t learn about an incident until long after it happened. When they did find out in time to take action, they often didn’t know how to respond to the incident.

We looked at these facts and figured it might help if conference organizers had an easy way to:

  • Educate attendees in advance that specific behaviors commonly believed to be okay (like groping, pornography in slides, etc.) are not acceptable at this conference.
  • Tell attendees how to report these behaviors if they see them, and assure them they will be treated respectfully if they do so.
  • Have established, documented procedures for how the conference staff will respond to these reports.

But we knew that conference organizers are very busy people, and very few of them had the time to write something like this. We figured that if we wrote an example policy that could be easily adapted to their needs, we could save them a lot of time and energy, and reduce harassment at conferences at the same time.

One year later, it looks like we had the right idea! Now it’s almost easier to attend a open tech/culture conference with a policy than one without. The response has been overwhelmingly positive, from attendees of all genders to speakers to organizers, and especially conference sponsors. Sponsors like any way to reduce the chance that their name will be associated with bad press.

You can help encourage adoption

Our goal is to make policies like this obsolete because everyone knows how to go to a conference without ruining it for the people around them. But we’re clearly not there yet, as this incident from October 2011 shows. One way you can help change the culture of open technology and culture is by encouraging the adoption of a similar policy by the conferences you attend.

Here are some of the common arguments against adopting a policy that addresses the three points we describe above.

This has/will never happen at my conference!

Congratulations! Some conferences are small enough or exclusive enough that it’s easy to end up with a group of people who all agree about appropriate conference behavior. Generally speaking though, as a conference gets larger or easier to attend, the mathematical probability of someone with significantly different ideas attending the conference increases until it is a near-certainty.

Next, if you believe there’s never been harassment at your conference, you might want to do a little asking around. If you don’t have a well-publicized method to contact the organizers about harassment at the conference, you’re unlikely to hear about it. When this policy was first posted, many organizers went back and asked attendees if they’d ever heard of harassment at previous conferences they had run and found the answer to be yes surprisingly often.

Finally, a great way to keep up a perfect record of no harassment is to adopt a policy that tells attendees you expect them not to harass each other.

Listing specific behaviors is unnecessary/insulting/ineffective/negative/etc.

Unfortunately, the overwhelming evidence from previous incidents shows that many of the people involved had absolutely no idea that what they were doing was unacceptable – and in fact were quite angry to discover that there were some unspoken rules that no one told them about. You may not enjoy telling people the rules specifically, but people hate breaking rules unknowingly even more.

To be blunt, a non-trivial percentage of speakers at open tech/culture conferences view pornography in their slides as simply good speaking technique. Telling them, e.g., to only include material suitable for a diverse audience won’t change their behavior because they believe everyone enjoys a little pornography in their technical talk. The only way they are going to stop including pornography in their slides is if you tell them not to, in so many words. Another non-trivial percentage believe it’s perfectly acceptable for a man to touch a woman on any part of her body without her consent if either the man or the woman is drinking alcohol. They believe this is appropriate behavior, so asking them to, e.g., “Be respectful of other people” is not specific enough to change their behavior.

This policy will hamper free speech and ruin my talk!

Conferences and their topics vary, but we have yet to attend a conference in open technology and culture in which a talk required the harassment of attendees in order to get information across. We’re not sure, but we suspect you can, e.g., teach people about file system semantics and keep the audience’s attention without employing sexist jokes. (I’ve done it more than once!)

Conferences in which talks about sexuality, racism, etc. are on-topic are encouraged to add exceptions for these talks and give guidelines on talking about the subject while respecting the attendees. We encourage them to send us their modifications so we can add them to the options in the example policy. Here is one example of the policy as applied to a talk about sexuality by Cindy Gallop at the Open Video Conference 2011.

In the end, you can always vote with your feet – you can preferentially attend, speak at, and help organize conferences with policies against harassment.

A note: We want to explicitly acknowledge the fact that harassment at conferences is not just a problem for women; in fact, we’ve heard many reports of men being the target of harassment, or being disgusted or creeped out by other attendees’ behavior. In this as in many cases, the changes that make open technology and culture more welcoming (and safer) for women are the same ones that make it more welcoming for everyone.

Conference speakers: Support anti-harassment policies in your speaker proposals

This post originally appeared on the Geek Feminism blog and is reposted here for posterity.

The Linux.conf.au 2012 proposal deadline is in a few hours, which gives you plenty of time to cut and paste the following into your speaker proposal:

I believe conferences should provide a safe, harassment-free environment for everyone. I ask $CONFERENCE to officially adopt and enforce a code of conduct or policy for attendee behavior that specifically forbids known problem behaviors such as pornography in public spaces, sexual harassment, and bullying.

If $CONFERENCE does not have a policy in place by the speaker notification deadline, I must regretfully decline any invitation to speak.

For more information, see:


Conferences value their speakers’ opinions greatly and listen when they speak. Donna Benjamin showed this when she organized a demonstration of support for a code of conduct at OSCON 2011; at least nine speakers in favor of the proposal edited their official OSCON speaker biographies to include a statement of support.

Friendly conference update

This post originally appeared on the Geek Feminism blog and is reposted here for posterity.

We announced a generic conference anti-harassment policy a couple of weeks ago. Since then several conferences have adopted anti-harassment policies, including Linux.conf.au 2011, FSF’s LibrePlanet 2011, and now all of Linux Foundation’s events have an official anti-harassment / discrimination policy. This includes 8 events in 2011 alone, including LinuxCon North America, LF End User Summit, and Kernel Summit.

Those of us who have attended Linux Foundation events will probably agree that their policy simply puts into writing what they were already doing. Other organizations which already have strong agreement on both standards of behavior and internal decision-making may be interested in adopting Linux Foundation’s simpler, streamlined policy. It is short enough to quote in its entirety here:

Linux Foundation events are working conferences intended for professional networking and collaboration in the Linux community. Attendees are expected to behave according to professional standards and in accordance with their employer’s policies on appropriate workplace behavior.

While at Linux Foundation events or related social networking opportunities, attendees should not engage in discriminatory or offensive speech or actions regarding gender, sexuality, race, or religion. Speakers should be especially aware of these concerns.

The Linux Foundation does not condone any statements by speakers contrary to these standards. The Linux Foundation reserves the right to deny entrance to any individual.

Please bring any concerns to to the immediate attention of Linux Foundation event staff, or contact Amanda McPherson, Vice President of Marketing at amanda (at) linuxfoundation (dot) org. We thank our attendees for their help in keeping Linux Foundation events professional, welcoming, and friendly.

(I helped write this policy as part of my pledge to help conferences adopt anti-harassment policies.)

Conferences that already had official harassment policies at the time of that announcement include OSDC and Ohio LinuxFest (one of the sources for the generic policy). LCA 2010 also deserves credit for including a clause on discrimination in its terms and conditions.

If your conference has an anti-harassment policy, let us know and we’ll blog about it on Geek Feminism! You can also add it to the list of conferences with an anti-harassment policy. If you are going to a conference that does not yet have an anti-harassment policy, and you would like to help change that, check out our list of conference organizer contact info.

Back on the speaker circuit

Over the years I have attended a lot of conferences, as you can see from this photograph of my conference badge collection:

Table full of badges

The badges are sorted into columns, starting with 1995 – 2000 on the left and 2009-2010 on the right. If the table was bigger and I didn’t have to squish the years together, you could see there was a big bump around 2006 – 2007 and then it leveled off.

I cut down on my conference attendance for many reasons, most of which were personal and most of which no longer apply. But I would like to start attending more conferences, and I would also like to promote the new open source conference anti-harassment policy. So I came up with a publicity stunt.

I swear an oath upon the K&R C Programming Manual that, for the entire year of 2011, if your conference, fest, users group, or other open-source related gathering:

  • Adopts an anti-harassment policy based on this or similar
  • Arranges some method to cover my expenses
  • Does not conflict with a prior engagement of mine

I will:

  • Speak at your conference about a technical topic and/or women in open source
  • Help you write your anti-harassment policy
  • If I can’t attend due to conflicts, help you find another speaker

Or, if you prefer, I will promise NOT to come to your conference – whatever floats your boat.

Some of the topics I can speak on are: Linux file systems, the TCP/IP Drinking Game, careers in open source, women in open source, btrfs, kernel development using Usermode Linux, practical git usage (ugh), kernel profiling and tuning, practical kernel development tips and tricks, and a variety of obscure and rather boring file system related topics (union mounts, relatime, chunkfs, making fsck faster, parallelizing IO, making ext4 64-bit, etc.).

I think I’m a middling good presenter – I got invited to speak at LCA (known for high standards on speaker technique), I got invited back to speak at LCA, and while people laugh during my talks I think they are usually laughing with me. Or at least that’s what they tell me.