I'm definitely doing something wrong:
Dalia and I (and a number of others!) had a conversation on our first Twitter Spaces session about "Comments: Good or Bad?". I argued for minimising the comments we have in our code, and in this blog post I want to explore how to do this in more detail.
Guess what? A conversation I was having on Twitter got me excited enough to blog again. Perhaps reading Twitter isn't such a depressing waste of time after all.
Continue reading "Who’s “Allowed” To Review Code?"
The problem with code reviews is that it is too easy to end to end up with code targeted at the lowest common denominator of the ability of the reviewers.
— Ihe Onwuka (@mazihe) October 21, 2020
Now I've been speaking at (mostly Java) conferences for a while (six years now), I get asked to present at a lot of conferences. Obviously all these conferences are mostly interested in my terribly educational talks, but it's also because I'm a technical woman and there aren't very many technical women speaking at conferences.
In my experience, conferences want to do the right thing - they want a diverse line up of speakers, they want to attract diverse attendees. Often this is not as easy as it may seem, and frequently conferences are Twitter-shamed for not having enough women speakers. When it gets to this point (and often before), conferences frequently ask me for advice on speakers they could invite, and how to attract more women.
So, I wrote a long email to the London Java Community in answer to an excellent question: "What can men do to support Women in Technology?".
It's a bit of a brain dump, by no means comprehensive, and is in answer to a specific question in a specific context, but I've been asked to make the information public so it can be useful in a broader context. So here it is.
I was asked three simple questions about my thoughts on Java turning 20, and ended up writing a guest post for Voxxed. No wonder I can’t seem to find the time to write these days.
Since publishing yesterday’s post, I’ve had a lot of great comments, so I thought I’d write yet another post to answer them.
Firstly, let me state that I don’t think blind CFPs are a complete waste of time - for example, I know that there are circles in which speakers will refuse to submit if the selection process isn’t blind. But what blind CFPs are not, is an answer to “How do I get more women speakers at my conference?", especially if the problem is not that the women are being rejected, but that they simply aren’t submitting talks - this is the situation I see most frequently in the conferences I speak at/help organise.
You’re missing the point of blind CFPs
I want to address the point “Blind CFPs are to protect the organisers, not simply to improve diversity”. OK, great. So in the case of the conferences that had only two women out of 200 submissions: if you make that process blind, whether you select both women or not, when there are nasty comments about your lack of women speakers, you’re protected - you did the right things, you had a blind CFP, and just not many women made it through that process. Women simply aren’t writing compelling abstracts and/or aren’t picking interesting subjects and/or clearly aren’t interested in presenting at conferences.
Protect your arse if you like, that’s fine, I see the attacks against organisers and I understand the need to be able to say you did the right thing. But notice that once you’ve protected yourselves, the blame is back on the minorities for not meeting your criteria.
If the purpose of a blind CFP is to protect the selectors, then let’s not pretend it’s an answer to increasing diversity of speakers, it’s a piece in a larger puzzle.
An abstract is not a representation of your talk
I don’t think I can emphasise this enough: blind CFPs have a very big weakness - there is no direct correlation between a good abstract and a good talk. I’ve been watching a bunch of talks online lately, and I watched one with an abstract that absolutely called to me - here was the missing piece for a technology that I simply can’t get my head around. I gave an hour of my life to that talk, each minute I was wondering when this person would get to the bits mentioned in the abstract, so I watched it all the way to the end. They never delivered what the abstract promised, and their meandering style meant I learnt almost nothing from the talk. That’s an hour of my life I’ll never get back. Good thing I was playing computer games as well at the time, if I’d paid money for that I would have been very peeved.
There’s another weakness in this process - if you go to a number of conferences (like I do) you kinda learn which speakers are doing which talks. If I submit my “AngularJS, HTML5, Groovy, Java and MongoDB all together - what could possibly go wrong??” talk to a blind process, if there’s anyone on the selection committee who’s even vaguely involved in the Java or MongoDB spaces, they’ll know that’s me. And they can Google the talk and decide whether they like it (or with any luck, the buzz-word crammed title will do the job and they’ll say yes, although probably AngularJS is so 2014).
An abstract is like your presentation in the same way that your CV is like your ability to do your job - there’s a skill to writing good abstracts (or CVs), but mastery of that skill does not necessarily mean mastery of the thing it represents - the ability to present or the ability to do the job. I’m not going to be sold on a blind process being a good way to build “the best content available for the conference” until this disconnect is somehow solved.
For goodness sake Trisha, the postscript is nearly longer than the original post
I’ve listened to feedback from a bunch of people, and I have learnt more things (thanks everyone!). I am more sold on a blind CFP being maybe part of a bigger process. But I do not believe it fixes the “we need more women speakers” problem, which was the original claim that I wanted to dispute. Conferences have to be much more active to address this issue.
And I am not sold on an abstract being a fair representation of the quality of a talk. Maybe a more realistic “blind” audition would be to give the talk to an audience, like a user group, who rated the talk: the organisers wouldn’t watch the talk, but they could use the ratings to decide. Yes, the user group could also have biases that affect their rating, but when the speaker gives the talk at the conference, those attendees will also have their biases too.
Blindness is not a realistic option, we just have to keep putting women (and other non-white-straight-men) in front of audiences until it’s no longer a novelty.
Off the back of yesterday’s post, I received a number of comments and questions around blind CFPs (Call For Papers - usually to get into a conference you submit to a CFP) for conferences. I often hear it said that a blind CFP will fix, or at least improve, the diversity imbalance at conferences.
I don’t believe this.
Firstly, I should caveat this with: this is going to be a blog post based on my experiences and chats with people, and I have zero science to back this up. Sorry. But I do have some numbers.
There are a couple of London conferences that I won’t name (because I haven’t asked their permission not because they’ve done anything wrong) that in some years have received only 2 from around 200 (or more) submissions from women. These conferences have accepted 100% of the submissions from women, and then been open to criticism for not having enough women speakers.
A blind CFP to remove unconscious bias from the selection committee might have resulted in zero women speakers.
Now I know the next argument is that advertising a blind CFP might encourage a wider diversity of speakers to apply. This does seem to work in some parts of the industry (I’ve been talking to women from tech scenes that don’t intersect with mine and I’m starting to see that we have different problems in different segments of the wider development world). But for my world, which is largely Java, with some forays into Agile, NoSQL, Open Source, and Enterprise, the women I speak to aren’t scared they’ll be discriminated against if they put their name on a CFP. Besides, they don’t have anything to lose by submitting and being rejected. They’re more afraid they’ll be accepted, and then they’ll have to give a presentation! No, the women I’ve spoken to in my world don’t submit because they don’t feel ready, they don’t think they have anything to talk about, and they’re worried about the sorts of questions they’ll get asked when they’re standing up there. Those fears aren’t allayed by having a blind CFP.
If there are any fears of being judged as a woman, it’s when we’re standing up in front of people at the conference, and no amount of blindness during the CFP process is going to make us feel better about that.
For me, personally, I find the idea of a blind CFP terrifying:
- Firstly, if you have such a problem with judging people based on race or gender that you have to anonymise the selection process, I’m not sure I want to speak at your conference where my gender will be blatantly on show.
- Secondly, when I was starting out as a speaker, I was new to writing titles and abstracts and expected that they would not be good enough to get in. With a blind CFP I would never find out why as I could never ask for feedback on my anonymous abstract. As a novice speaker, I expected my gender would help me be selected, not hinder.
- Thirdly, as an experienced speaker I know my abstract-writing has not really improved, but I know my presenting has, and I know there are videos of me out there doing a good job as a presenter, this is my audition tape.
When I was starting out as a speaker, I knew there weren’t enough women presenters around. And although I hate the idea of positive discrimination and hate the idea that men might think I got into my position merely because I’m female, I knew that being a woman made me stand out in all those submissions. And I knew that enough conferences would take a chance on me, if only to improve their diversity, and that’s all I needed to bootstrap my speaking career - once I had spoken at enough conferences, once I had videos of me doing that, I would find it much easier to get into the next conferences. And that’s worked. And now, as an experienced (female) speaker, I fear blind CFPs as they won’t know who I am or what I’m capable of.
So that’s my opinion as a female speaker. But I’ve also been fortunate enough to be on the programme committee for a couple of conferences, and after three years or so of speaking at conferences I know a lot of conference organisers. Let’s assume you’re “merely” trying to build the best conference you can around your subject whilst not discriminating against anyone (i.e. let’s assume you’re not going out of your way to actively promote diversity). How can you do this when you only have abstracts and titles to go on?
I am aware of the studies around blind auditions for orchestras. And I understand how tempting it is to try to apply that to our world, since it has had an effect on the number of women in orchestras (but let’s not get too excited here, we’re talking about an increase from 5% to 25% over thirty years…). I think blind auditions would work really well for coding - we can look at just the code, without knowing anything about the physical body behind the code, and figure out if that person is good enough for us. We could even do blind pair programming, using screen sharing and chat (actually that’s a great idea, I think I’ll patent that!). Because the thing you’ve anonymised (the code) is the output you really care about in a programming job. If you communicate via chat you can even measure someone’s communication skills without judging them physically, another vital skill for being part of a team.
But a talk abstract is not the output we’re looking for. Sure, a good title and abstract will attract attendees to the conference and bums on seats in the actual talk. But the output you’re looking for is the presenter - their style, their approach to the topic, their ability to hold attention, their ability to impart information. We’re not always looking for the same things either, an interesting topic presented nervously but logically is fine, as is an entertaining presenter reminding us of fundamentals we already know (and any other combination of strengths and weaknesses). An abstract can’t tell you how good the presentation is going to be. Evaluating a presentation on a blind abstract would be like evaluating those musicians on a couple of paragraphs explaining what they’re going to play and how they’re going to try to make you feel when they play it - the words could sound great, but the musician might not even be able to play the instrument.
The best audition you can have as a presenter is to, well, present. Videos, audio clips, even screencasts. But the problem with presenting is that it’s so freaking obvious that you’re a woman (or other variation on non-white-straight-male). If you can’t objectively evaluate a presentation without taking away the presenter’s physical body (for colour or gender or whatever), then you’ve got real problems. Your audience is going to be faced with that physical body for an hour of their life, and you can bet they’ll be judging them too. You want to be damned sure that you’ve selected a competent presenter, and you’ll need to stand by that choice. Saying “oh, well, they didn’t seem boring in their abstract” is not going to make the audience feel like you did a good job of the selection.
There’s another point about non-blind selection processes, other than the process allowing you to select on presentation ability rather than abstract-writing ability: you can be more active in promoting diversity. I am against “affirmative action”, but if the proportion of women submitting presentations is lower than the proportion of women in the field, there’s something missing. Here are some of the things you can do if you identify a lack of a certain demographic - as usual, I’m going to use “women”, but you could equally argue this for any demographic, even not-enough-C#-developers or too-many-startups-not-enough-enterprise:
- Reach out to relevant user groups and communities asking them to submit. In London for example, there’s Ladies Who Code, Women Who Code, Girl Geek Dinners, PyLadies, Rails Girls, Women in Data, Girl Geek Meetup, Women in Games Jobs, Girls in Tech, Girl Geeks Unite, and a whole host of women techies meetups … And don’t think I psychically know about all those because I have two X chromosomes, I found half of those just now, in less than two minutes, by Googling “Girl Geeks London” and “Women Tech London”
- Put the faces of the women you do want to have speak (or the women on your programme committee, or both) on your website - showing these women sends the message that this conference is for women. It should also impact the diversity of your attendees.
- Reach out to specific women speakers and ask them to present. I’ve talked about this before, women speakers are few and far between, and a) they get booked in advance and b) are busy enough not to have to (or want to) submit via a CFP process.
- There are more ideas here
And there’s more - you don’t just have to drain the pool dry of known competent speakers-who-tick-some-diversity-box. With a non-blind process you can actively improve the submissions you do have, assuming you’re not one of the conferences suffering from a complete lack of submissions from those people. For example, if you’re looking to increase the number of women presenting at your conference, if you have women submitting but they don’t quite meet your criteria (maybe the abstract is weak, the title isn’t catchy, or maybe the presenter simply isn’t quite ready yet) you, as the conference organiser, are in the perfect position to mentor these people: you can help them re-write their abstracts; you can offer them mentoring and training for presenting. If there’s no record of them presenting but you’re interested in evaluating their ability, you can set them up with a user group to present at - as a conference organiser, you have the network contacts that new speakers might not. You can run sessions aimed at people who have never done public speaking (I’ve run a session like this four times now, at the LJC and at Ladies Who Code - I had over 70 RSVPs for the last London Ladies Who Code session, twice the number of the last LJC session, so there are plenty of women interested in breaking into this field, who are looking for training, who don’t feel ready to present yet).
So, what are my points?
A blind process of selection can be applied to something where the output you’re selecting for can be effectively anonymised: music auditions, code, technical writing, etc.
A blind process where you select based on criteria that are not the end result will only be effective if it’s the selection process is biased, or if submissions to the selection process are lower from discriminated-against-groups because they fear a biased selection process.
For those conferences that are getting very tiny numbers of submissions from women (I’m picking women as the example but you can pick any group you want), a blind process could lead to even fewer selected women speakers. Instead of switching to a blind process, you’re probably going to want to check:
- the messaging around your conference (remember masculine words like rock star, ninja, competitive, hardcore, challenging etc put women off applying to jobs, the same probably applies to technical conferences)
- subliminal messages on your website/info (one of the conferences I mentioned, for example, sent out a couple of emails containing photos of previous years’ conferences which happened to only show men - this was not even representative of the conference, but it can have an unexpectedly big effect on the subconscious)
- have you showcased the diversity of the conference, or previous years’ conferences, with photos, interviews, testamonials?
A non-blind submission process allows you to be more active in promoting diversity - instead of hoping that the submissions are representative of the diversity in your community, and that selecting the most interesting talks will naturally lead to improved diversity, with a non-blind process you can track exactly how many women are submitting, address problems in this area early and offer mentoring to those who aren’t quite ready yet. Remember that women only apply for jobs they feel 100% qualified for and take fewer chances, they might need that extra push.
So, no. I don’t believe that implementing a blind CFP will address the lack of diversity in conference speakers. I know it helps in some circles, but I personally think that conference selection committees should a) be much better at identifying their own biases and actively attempting to overcome them b) be active in improving the diversity of their conference, and c) be offering feedback and mentoring to those who are not quite ready yet in order to improve the overall ecosystem.
Diversity isn’t going to accidentally happen when we stop looking at gender and race and other visible you-don’t-belong-here cues. Diversity is only going to happen if people fight for it, and if we actively train, mentor, and support those trying to break through the barriers.
I wrote a follow up based off responses to this post, covering additional questions raised.
Last month at Devoxx I was in a session discussing what we can do to encourage more diversity in our speakers (specifically, although not limited to, increasing the number of women speakers).
I’m going to outline the things I remember being discussed, although as usual we did not find the answer to the problem, only identify some issues and explore some options.
This is a very chaotic blog post, because if I don’t post it now I’ll never post it, and it’s better if my thoughts are scrawled down and posted than if this all goes to die in my drafts folder.
Firstly, what stops people from stepping up to speak?
I don’t have anything to say
This is the very first thing any developer says when you ask them why they haven’t thought about presenting at conferences or user groups. Turns out, you always have something to say. For example, our most popular sessions at the Sevilla JUG and MUG are “Introduction to…". You hear from plenty of experts who have travelled the whole road to some pinnacle of excellence, but what many developers need is simply someone a couple of steps ahead of them to lead them forwards.
You always know more than someone, you don’t need to know everything, and you don’t need to know more than everyone
Why should I be role model?
So many women feel this, and I’m completely sympathetic to this view point: I’m a woman developer, getting along, happy in my job, happy being a programmer, why the hell should I have to stand up and present and blog and be a role model for other women? Men don’t have to do that. Men can just be developers.
In addition, there’s the fear that if you do stand up, become a role model, that somehow people will think you’re representing all women developers. You’ll be worried because you can only represent who you are, what you do, you can’t represent a whole gender in a single presentation or a single interview. And the pressure! What if I mess up? What if I say something my gender does not agree with? What if I say something that causes my whole gender to be judged?
Totally understandable fears. Totally true that many people will ask you to be the woman developer, to answer for all women. Well, that’s why I’m asking more women to stand up, to be role models - if we all step up, it will no longer just be me, and you, and that other one over there. There’ll be more of us, more differing voices, we’ll stand out less. But we have to stand up first. And in the meantime it’s us and our points of view they’re listening to. Wouldn’t you rather voice your opinion than have someone else speak on your behalf? Especially when very few people really can represent you, who you are and what you want.
What if I can’t answer the questions?
You’ll be surprised at what you do know, so people who present run into this less than you’d think, even if you’re just beginning, but just in case here are some tips:
- Repeat the question, paraphrasing. Firstly, this ensures you’ve actually heard the question correctly. Secondly, it allows the whole audience to hear it, and thirdly, but most importantly, it buys you time to think.
- Be honest - say you don’t know, and move on. If you know somewhere that might have the answer, point the person to that resource. But a technical audience actually values the answer “I don’t know” more than you think. Especially the person who asked the question, it makes them feel really smart!
- If you don’t know, but you think the audience might know, throw it back to them. Either to the person who asked the question, if that feels appropriate (especially if this is one of those “look how smart I am” questions), or to the audience as a whole. In conferences in particular, there’s a good chance someone there knows the answer.
- If you’re genuinely terrified of all questions, say at the start of the presentation that you won’t take questions in the talk. Offer an alternative, like via Twitter or e-mail, if you can.
What can conference organisers do?
Specifically invite the people you want presenting
Don’t just wait for them to submit. If you want more women, more diversity, invite them. And invite them early…
Secure your women presenters early.
Recognise that there are relatively few women presenters on “the circuit”, and understand that everyone wants them to speak at their conference. This rare breed, the female technical presenter, gets booked up well in advance. They’re probably as in-demand as your keynote speakers, and can probably devote less time to appearing at conferences than those who have made it to keynote-standard (I’m making a sweeping generalisation that your techy women have day-jobs as developers and your keynote speakers are either professional presenters or at least can argue presenting as a critical part of their job).
So, you have a limited resource (experienced women presenters) that’s in high demand. You need to reach out directly to them, invite them, and secure them well in advance of your conference. Many big conferences, for example JavaOne and QCon, will invite speakers six months in advance. If you give women speakers a month’s notice before your conference or event, they simply are not going to have the space in their schedule at that late notice.
Showcase your diversity on your website
If you haven’t already heard of microaggressions, go and read what that means before you get any further. There’s a very simple thing you need to do with your website, whether you’ve opened for CFP or you’re trying to get attendees to register - make sure the photos on your website show the type of diversity you want to encourage. Only yesterday I was looking at the website of a big conference, who I know for a fact are working really hard to ensure diversity, not just with gender but race, age, and where possible in less physically obvious dimensions. The conference is many months off, so the only photo they have on the website is a background photo of the exhibition hall from some previous year, crammed full of attendees - white, male attendees, as far as it’s possible to tell. It’s extremely subtle. You might not even notice, as the photo is behind the headers and text of the site. But I noticed. And you can be sure the subconscious of every person who went to the website to either submit a proposal for a talk, or is considering asking their employer to pay for the registration, also noticed. The conscious mind might be able to shrug it off. But the subconscious of a non-white and/or non-male person notices and plants the subtle seed of “this conference is not for me”.
So, conference and event organisers - if you want more women applying to speak at your conference, show more women on your website. If you want a more diverse audience, show more diverse faces. If you’ve invited your women presenters early and secured them early, show their photos on the site - this sends a strong message to other women that their talk proposals will be taken seriously.
Run warm up nights to try out new speakers
I know conference organisers get lots of submissions, and I know it’s hard to tell which are going to be good speakers, especially if they’re new speakers or don’t have a video. If you’re genuine about improving the diversity of your speakers, run free nights to a) promote your conference and b) try out new speakers. Probably a good idea to have a new speaker and a known speaker to attract the crowd and give newbies a chance.
Don’t be so damn picky
- New speakers need to be encouraged. Find a way to see their content without just discarding an unknown name.
- Existing speakers might not have the luxury of creating new content. Many conferences want genuine practitioners, not just evangelists, but those people have day jobs and can’t be creating a new talk just for your conference.
Leave space for new speakers
…and don’t just have 7 slots for a pro speaker
Look at your swag
Not just guys t-shirts. One conference gave out cupcakes for example. But it doesn’t just have to be handbags and high-heeled shoes, just take a quick look and see if your swag is generally useful for techies, or if it leans towards men only. Men only swag (men’s t-shirts, bottle openers, subtle things like that) subliminally says “you don’t belong here” to the women.
You can’t tell every individual why they were rejected, but you could look at all the rejected submissions and give broad indications of the type of CFPs that didn’t make it. Maybe write a blog post explaining which sorts of talks didn’t make it and why. This will help speakers create content that better fits your conference.
Be specific about what you’re looking for
If you have tracks you’re finding content for, put them on the website/CFP, so people know what the topics are. If you have an idea that you want 30% beginner material and 30% very advanced, make that clear too. Especially if you want only beginner or only advanced. People will submit more appropriate talks.
Have a beginner’s track
…not just for the speakers, for the content too. Beginner speakers may be more comfortable presenting “Introduction to…” talks.
Having a beginners track, whether it’s for content or speakers, will make newbies feel more included.
What other things can we try?
User groups can grow talent
- Practice writing CFPs
- Help each other
- Presentation skills
- Lightning talks, book club, different meetup styles (like fishbowl)
- Run an open conference to encourage new speakers
Women’s user groups are instrumental
So here’s something interesting. I’m sceptical of all-women user groups as I feel it separates us from the communities we should be contributing to. But women’s user groups can help us to integrate better. For example, Duchess France helped and encouraged its members to submit talks to Devoxx, gave feedback on CFPs before they were submitted, and provided a safe place for the women to practice presenting beforehand. I think this is exactly the sort of thing women’s user groups could be fantastic for.
Advice for budding speakers
- Just Do It
- The more you do it, the more comfortable you’ll be
- If you’re a speaker, it can make you feel less of an impostor at a conference - you’re not just a woman developer, you’re a speaker!
- Everyone’s afraid, it’s not just you
- Every developer wants to learn
- The audience is on your side
- Start Small (user groups)
- Be technical. But you don’t have to be super smart
- Bootstrapping is hard - you haven’t presented so you can’t get into conferences. Again, user groups can help with this.
- Pair. Ask a more experienced presenter if you can co-present with them.
- Ask for mentors
- Ask for feedback
- “I don’t know” is perfectly acceptable.
I appreciate there is so much more I could have covered, and many topics deserve more depth. But this is the brain dump from the Devoxx session, and maybe one day things will be expanded into their own topics.