First timer thoughts from FOSDEM15
Last weekend I went to FOSDEM, and Brussels, for the first time. Normally my January free software conference of choice is linux.conf.au, but being in the northern hemisphere for a change afforded me a new option. A free-as-in-beer conf with over 500 speakers and 8,000–10,000 attendees is free software on a whole new scale, for me.
Two themes I saw again and again were the importance of usability, and excitement about declarative programming.
Talks
Hopefully there will be videos to come…
On Saturday, my first talk up was Identity Crisis: Are we who we say we are? by Karen Sandler (interview). She talked about conflicts of interest, the slippery nature of “we” in free software communities, and her work with the Software Freedom Conservancy. SF Conservancy, to my understanding, plays a similar role as the Apache Foundation, the Python Software Foundation, or the GNOME Foundation would play, but for any free software project that wants a legal layer (e.g. for defending license or trademark problems) without the hassle of setting up their own organisation. Projects like Git and PyPy have come under their umbrella, and I imagine a lot more are likely to take it up in the future
She also announced the launch of Outreachy — the rebranding of GNOME Outreach Program for Women. Both the dropping of women and GNOME are significant — the program has long since expanded beyond GNOME, and now the name reflects that women are not the only under-represented group that the program aims to reach.
I first saw Karen talk at LCA2012 where she talked about becoming a cyborg lawyer — a great talk if you’ve never seen her speak about this topic.
I would have attended the Python room but it was packed every time I walked past so…
Crypto Wars 2.0 and the Free Software Response by Aaron Williamson was really really excellent. It was a summary of US laws since the 1990s related to cryptography and surveillance. Aaron pointed out that the NSA is issuing the same warnings now about the need for lawmaker’s access to software/devices, that it made 15 years ago. But why did they not push the agenda 15 years ago? Because they had a “spyer’s market”, with programs like PRISM allowing access to many communications even without built-in backdoors. Meanwhile public outrage over the Snowden revelations has led those large tech companies to actually act to bring us closer to end-to-end encryption. Aaron then talked about how we are now reaching an opportunity for free software, because as outrage wanes, the efforts of those companies on our behalf may also wane. He made it clear:
“This is not a technical problem — it is organizational.”
We have the tech, where’s the adoption? Stuck behind a complex install process.
It’s hardly news to say that free software often suffers from usability problems, but Aaron had a further analysis that I found very interesting. He argued that we have essentially taken ESR’s description of the ‘bazaar’ model of development to heart, but its flaws are now holding us back. For example with enough eyes all bugs are shallow — few would posit that so convincingly these days, post-Heartbleed. The idea of user-developers happily scratching their own itches also falls down around these kind of critical projects, which for many years had a bus factor of 1. There have been many efforts to redress these errors now, such as the Linux Foundation’s Core Infrastructure Initiative. GPG funding was also recently announced. Finally the third arm, meritocracy — well, does that still need debunking? (Probably—see discussion under CoC section.)
I like the idea that we are ready for a new paradigm
to describe free/open source software development.
It also reflects the reality I see in the Python world, which has managed to blend corporate and community contributions quite well. As have others. But it has taken some convincing that corporations that rely heavily on Python, have a duty to “contribute back” in the form of paying people to work on the core of the Python language and its libraries.
In the afternoon there was an informal pytest meetup, with 6 of us including Holger Krekel who founded the project. It was a dense discussion with a mix of philosophical questions, high level planning and at times detailed specifics of the inner workings of Mercurial and Bitbucket.
On Sunday, I started with The Story of Rust by Steve Klabnik (interview). Rust is a new systems programming language developed within Mozilla that will have its 1.0 release sometime this year (2015). Steve described Rust as offering something for three distinct camps:
low-level access and speed for the C crowd,
toolset and development ‘feel’ for the scripting language crowd,
and idioms and safety for the functional programming crowd.
I like this mix a lot! Since arriving at Python I enjoyed it enough, and have not had problems hard enough, to spend any further time looking at lower level languages, but I remain aware that in the eyes of some that makes me a lesser developer. Rust is the first systems programming language I have heard of that actually makes me want to take a stab at it.
At the moment, given Rust is still pre-release, it can be hard to find non-toy examples to try it out in. Except of course for R&D Mozilla project Servo, which is where this talk may be useful: Servo (the parallel web browser) and YOU! A beginner’s guide to contributing to Servo by Josh Matthews.
I went to see Get ready to party! by Larry Wall (interview), mostly out of a sense that you should take your chances to see a free software legend in person. He is obviously very beloved — he was the only speaker I saw to get a standing ovation. Not being well versed in Perl history or Tolkien, many of the jokes in his talk went over my head, but I appreciated his sense of humour and play that were amply on display.
Finally I saw Can Distros Make the Link? Let’s Package the Customizable, Free Software Web of the Future! by Deb Nicholson and Chris Webber of MediaGoblin. This was in the Distributions room, and was a kind of open request/plea for distro packaging people to think about how they could experiment with ways to make installing complex web apps simpler, for people who want to self-host their email, or website, or microblog, or Flickr/YouTube clone (a la MediaGoblin). Perhaps with another “layer” distro package management? Hopefully it will kick-start further ideas and discussion. Sandstorm came up a lot but it is still very early days. Even something as seemingly well “solved” as distro packaging may take on a new life.
Code of Conduct, and culture
Well. I like going to community-run free software conferences. I like being exposed to lots of things I’ve never heard of, and having a hundred new ideas an hour. I like talking geek with people who care about the same projects I do. On the other hand, I don’t like the very in-my-face reminder of being a gender minority. And I really don’t like the reminder that many women avoid these events, avoid this community, avoid this industry, because they face exclusion and hostility, ranging from inconsiderate jokes to sexual assault.
I am happy that the lack of proper CoC was such a talking point that the organisers were moved to respond. To me the most important thing about a CoC is that it is a signal to potential victims of harassment. It’s not going to magically eradicate all problems, even if we could kick out every abuser, every “bad person”, which we can’t. (Note, not all acts are committed by “bad people” — and this black-and-white thinking may stop you from acknowledging something dangerous enacted by yourself or a friend or hero.) But the CoC should let potential victims know that they will be, or perhaps are more likely to be, believed and supported. They need to know: how will the organisers respond if a problem is reported? Thus, the first most important thing about a CoC is that the organising team believes in it…having any specific string of words is worthless unless the team is prepared to stand behind it.
While I heard estimates of 10–15% of attendees being women,
to me it felt much lower, maybe 5%.
The conversations around the CoC topic surprised me — they felt “so 5 years ago”. So did attending a conference with such a low percentage of women, to be honest. It does surprise me that there are many people who probably have an opinion on the following topics:
- systemd debate
- Node.js/io.js fork
- Rust vs Go
- Docker
…yet the arguments they put forward about the merits of having a CoC are c2010. Butif you care about free software having a fighting chance of being able to change the world, efforts to make a more welcoming environment for women and other minority contributors are easily more impactful than all of the above. So seriously, don’t be showing your surprised face. Ashe Dryden has written a wonderful Code of Conduct FAQ.
Another point is that a welcoming environment is not created solely by a CoC, it can be created by every attendee speaking up where they see inappropriate behaviour or hear inappropriate comments. Karen Sandler had her audience practice this — a goofy collective “That’s not cool!”
But it worked — I heard an attendee call this out in a talk after an unfortunate ad-hoc pun. You don’t need to be the self-appointed fun police — you just need to voice your discomfort. I find it immensely comforting to hear a guy do this.
It’s kind of the best. Seriously. Try it sometime. Sparkle!
Jenny Wong organised a “Diversity and Inclusivity discussion” for Sunday afternoon — by which time the organisers had already tweeted their intentions to improve the CoC next year.
We will see!
