[diversity] Experiences from the Django community

Hi everyone,

I’m delighted to see that the RIPE community is taking important steps to improve diversity and inclusivity. As some of you might remember, I worked for RIPE NCC until 2011 and attended most RIPE meetings back then, though none since.

Since then I’ve been doing a lot of volunteer work around diversity and inclusivity, in particular in the Django community. Django is the most popular web framework for Python, developed as an open source project for the last 12 years. In recent times, the Django (and to a different extent, Python) community has made large steps in diversity and inclusivity. It is not uncommon for Django conferences to have an approximate 50/50 split between men and women on stage, even though talk selection is usually blind (i.e. without information identifying the speaker). There is much work left to do though.

I felt there may be resources, experiences and best practices that can be helpful for this group too, so I laid out a few things below. (I couldn’t reply to the existing thread, as I only just joined the list.)

A fundament is that the Django community and events have a Code of Conduct[2] with a strong process behind it, including things like tracking repeat offenders and (limited) sharing of information between conferences. I recently became a member of the CoC committee that manages part of this. For example, prior to accepting speakers, conferences often send their list of names so the committee can compare that to their list of past incidents. In the interest of transparency, the enforcement manual[3] along with most internal documentation[4] is public.

A CoC with a proper process is only a first step. Other efforts in the Django community include extensive outreach to marginalised groups, offering speaker mentors, financial aid programs, ensuring people moderate their alcohol consumption, accessibility information, live transcription, quiet rooms, childcare, and so on. As a side project, I maintain the Less Obvious Conference Checklist[1] which has a lot suggestions of this kind for conferences - most easy, some very hard - many of which can make an impact on inclusivity.

Some of the lessons I learned along the way that I thought may be relevant, reading your other discussions:

- D&I is more than “at this conference we claim you won’t be harassed or assaulted, or we will take action if it happens”. For many marginalised groups, experiences in tech and tech communities are so poor, that it requires building a lot more trust in that a community is actually doing better in being a nice and safe place, and not just saying so. Personally, I think building this trust is the single hardest issue.

- Related to this, if people aren’t reporting incidents under your CoC, it doesn’t mean they didn’t happen. People (including me) can be very reluctant to report, and if they don’t trust the process enough might just end up leaving and telling their friends to stay away. In Django events, we’ve noticed a small increase in minor incident reports - not because more are happening, but because we’ve built enough trust that people report them.

- Statistics can be helpful, for example in tracking whether there is progress compared to before. They can also give an indication of whether things you’ve done seem effective. However, I’m skeptic of having direct numbers as single absolute goals. On the other hand, you can probably say that if 2% of speakers are women, something isn’t going right somewhere.

On intersectionality: any D&I initiative that wants to stay away from the complexity of intersectionality is dead in the water, as is an initiative that wants to treat everyone equal. My position, even being part of some marginalised groups, and my ability to participate in a community and attend events, is incomparably privileged compared to (for example, on average) a trans woman of colour or someone with housing or income insecurity. We can not treat everyone the same, because we are doing this in a society where marginalised groups and many others are already not treated equal or provided equal opportunities. And simplifying efforts to “sex discrimination” is simply way too narrow.

I hope that some of this was useful input. I also do a talk at a number of conferences on empathy in tech, which touches upon D&I and how experiences from marginalised people can be so different, and why this matters so much, among other things. A video should be out soon. In August I’ll be doing a talk at DjangoCon US about the Django CoC, the processes behind it and some of the myths, which might be interesting to this group too.

Erik


PS: Could someone update the XX/XY labeling on the attendee gender graphs on RIPE labs[5]? Guessing gender from names is hard and inaccurate already, but sometimes all we have and I understand that. However, unless you have an algorithm that actually guesses their chromosomes, labeling the groups as XX and XY is wrong and completely erases transgender people.

[1] https://github.com/erikr/lessobviouschecklist
[2] https://djangoproject.com/conduct/
[3] https://www.djangoproject.com/conduct/enforcement-manual/
[4] https://github.com/django/code-of-conduct
[5] https://labs.ripe.net/Members/agowland/diversity-discussions-at-ripe-74