As I write this post, I'm sitting on the train home from OggCamp 2024.

It's been a good weekend - I've met some nice people, caught up with folks I've known for years, had some great food, listened to some great talks, and learned some cool new things.

Telegram chat

Something nice that the organisers did this year was to have a Telegram group for attendees - I found a few food recommendations, and also met a couple of people through it, and it was a nice lightweight alternative to i.e. Slack or Discord.

Food + drink

I ended up having some really good food and drink this weekend, so thought I'd recount it for the future.

I up with met Lorenzo for a drink and some food at the Wolf At the Door on Friday night, where I had a good Teriyaki Beef bao bun and some really nice korean fried popcorn chicken - which had a good bit of a kick!

It was nice to meet and have a chance to chat - albeit we both found it a bit loud, which I guess it was a Friday night? - but was nice to meet, chat Open Source, life and nice to have a new friendly face at the conference, and beyond!

On Saturday morning I had breakfast at Fiori Cafe which was a really good choice - and as I walked past it a number of times over the weekend, almost always had long queues waiting for a table! - and nice and quiet when I went. The shakshuka was really good - albeit it defeated me as it was very filling - and set me up really well for the day.

Per the Telegram chat, I saw a recommendation to Hello Oriental bakery which was also just round the corner and had a really nice Honey Roast Pork Char Sui for lunch.

In the evening, I met up with Lorna, Rob, Jon and a few others - new and familiar - for the world buffet, Tops which had some good bits, and was nice to chat some more.

This morning, I went over to Federal Cafe for a great eggs Benedict (with bacon) and a latte.

Today at lunch, I went back to Hello Oriental to try their small mix-and-match buns, and got another Honey Roast Pork Char Sui for an early dinner.

Computer science is in the way of climate science

There were some interesting points in this talk by Michael Winston Dales about the difficulties of building software for research, in particular fun bugs in the epsilon skew in large numbers of data points, the need to understand memory hierarchies, and that having an understanding of the key parts of your stack are incredibly important, as is versioning them, in case i.e. the GEDI telescope changes the algorithm for how they resolve certain things.

The mazy web she whirls: starting Open Web Advocacy

Stuart Langridge shared a look at the work he and others have done with the Open Web Advocacy group, and some of the impact they've had on making the Web a little more Open - or at least, stopped it become a little more closed.

On his blog, Stuart shares more information about the talk, including a recording from previous events he's given the talk at.

Annoying your loved ones with home automation

Charlie O'Hara shared some fun insights into home automation.

There were some good tips for annoying your loved ones including "make it more difficult than it needs to be" and "tape down light switches", as well as Charlie's exploration of her experience and learnings through the home automation journey.

This was a fun talk, and has made me reconsider if us putting time into Home Assistant is worth revisiting 🤔

How to organise a union at work

I unfortunately had to nip out half way through the talk, to prep for my own talk, but it was an interesting start on the process, and was followed-up with a panel on Sunday that I did attend.

There were some good notes around how different parts of the organisation may have different industrial power - for instance, engineering may have less power to make impact while working on a new product, as they can only really slow things down, but if SRE say they're not going to pick up on-call pages, that's a different level of impact.

Quantifying your reliance on Open Source software

I had a pretty packed room for my talk about dependency-management-data, which was great 👏🏼

I had a lot of interest after the session, with a few minutes of more in-depth discussion, and I was speaking to one of our party at dinner who was also going to be trying it out on Monday, which was very cool 🤓

A note to myself is that I should probably tweak the slides for a 25 minute talk, as there were a couple of bits I had to speed over a little more than I'd have wanted to.

Into the multiverse: a parallel universe where neurotypicals are the weird ones

Parul Singh's talk was incredible.

From my review of Parul's talk - also posted on JoindIn:

Parul's talk was incredible - her humour and reflection on her experiences of life as a neurodivergent person, through the lense of "what if neurotypicals were the weird ones" was really quite powerful.

I spent the talk very eagerly nodding along and laughing - a little distractedly - at her incredibly relatable take on the role reversal, and bangers like "Norman received feedback [that his colleagues felt weird about him because] he couldn't fidget".

I'm biased as someone who is neurodivergent myself, so related much more heavily to the talk, but feel that non-neurodivergent people in the audience could also understand the parallels.

I was very strongly reminded - in a great way - of Rachel Morgan-Trimmer's talk from OggCamp 2019 which was also a really powerful insight into life as a neurodivergent person, and I really enjoyed Parul's empathy-driven talk.

I especially loved the ending call to action, asking us all to consider what steps we could take to understand others in our life better.

My only complaint was that I wish the talk was recorded!

I unfortunately missed the first few minutes, as I'd had a few follow-up questions from my own talk (which I was very appreciative of) but thankfully was able to join the talk.

After this talk, there was a nice group of neurospicy folks all having a chat outside the theatre which was really nice, and I got to meet Richard, who joined our party for dinner later, which was nice.

This talk was absolutely the highlight of the conference, and I am very gutted for everyone who missed out!

Wrestling a Giant C++ desktop app into your browser

I was a couple of minutes late to the start of the talk, so I didn't realise until now that Michael Meeks was the CEO of Collabora Productivity, the company behind the web-hosted LibreOffice, and a contributor for many years.

This was a really interesting talk, and especially as it's been a while since I've been exposed to some in-depth Linux, it was nice to dive back in, in a very approachable talk that took us through the fork from OpenOffice, and taught us some of the incredible work that's been put in to improve stability, maintainability, learnings for putting it onto the Web, and the removal of a whole load of German code comments.

Lightning talks

I attended a couple of the lightning talks sessions across the weekend, and there was a great mix of topics and speakers.

89 things I know about Git commits (abridged version)

I put myself down to speak about my blog post of the same title, but trying to distill it to what was able to be spoken about in the 5 minutes I had, which as Liz called out, would've been one point ever 3.37s, which seems a little difficult.

I was glad to have got laughs out of the audience at the right points, and that there were a number nodding along.

I will say that I also found it really interesting re-reading my post.

In a very not-humble way, it's a really great post, and I was re-reading it while putting slides together for it thinking about how many good points and nuggets of insight I put into it. Not least because of all the contradictions in it, which were really fun to write.

I was so close to rewriting one of the slides ~90 seconds before I was ushered onto stage, but I resisted 🫣 I feel it may have flowed better, but not worth the additional adrenaline spike!

How healthy is your open source community?

Alan Pope (aka Popey) spoke about the Savannah Community Relationship Manager (CoRM) project, for better understanding your Open Source communities.

This was not a product I've heard of before, but it was super interesting to see how it could be used to better understand the community health of your project(s).

The most impressive feat was its support of many different sources of community discussion, such as Slack, Discord, Discourse, Salesforce, and GitHub and GitLab (where GitLab is usually lacking in support), as well as supporting an understanding of different repositories under a single "project".

We heard about how it can discover key members of the community i.e. based on how connected they are to different discussions and people, as well as helping you keep track of information about different contributors, learn how different folks are interacting with the project(s) and how engaged they are (in particular using the engagement pyramid - visitor -> participant -> contributor -> core).

As Popey mentioned, it's clear that its creator, Michael Hall, has been working on this for years and understands the key metrics and journeys that community managers want to understand.

There was a good discussion on possible privacy or consent issues related to turning this on across communities and different platforms, especially with how it may be able to connect identities across different platforms.

One of the audience members mentioned that due to their usage of Savannah, they'd found an issue that they weren't aware of - as there are often far too many things to keep abreast of - which was a fairly long-standing issue with a lot of community members looking for its implementation.

It looked like a really good product, and definitely something I'm going to consider looking into for oapi-codegen, not least because it could provide some insights into where we should really focus on issues with the most voices clamouring for support, as well as any other interesting things we can learn (aside from "we need to do more work").

Fighting Ens**ttification with ActivityPub

This talk from James Smith was a look at how you could improve your projects by removing the risk of enshittification by building in interoperability with the wider Fediverse.

James started by talking about some examples of the awful practices some products are following, making it so people wish they could be using something else.

We then learned all about ActivityPub, some of its common types, and digging into the under-the-hood things you need to understand to build on top of it.

We also dug into HTTP Signatures a little - which gave me some flashbacks from implementing a shared library for Java services at Capital One to secure our API-Gateway-to-service traffic - and how debugging with a lot of the ActivityPub/HTTP Signatures work can be difficult as you need to have servers talking to each other, over HTTPS on the public Web, which is a bit more involved.

James did mention Federails which aims to abstract a lot of this work for you, fitting nicely in the Rails-y way of doing things.

Let's talk about the Open Source AI definition

This was an interesting session facilitated by Paul Barker around the Open Source Initiative's work towards defining what "Open Source AI" actually means.

As someone who's mentioned they've not bought into the "AI hype", or even really played around with it that much, I know it is of interest to many, and is happening whether I'm bought into it yet or not, so I am tangentially interested in it.

Paul ran the session well with some good ground rules, and kicked us off with some good points, and then facilitated a good discussion, of which there were a few key areas of discussion:

• The Open Source Definition came as a follow-on to the Debian Free Software Guidelines, and as a way to standardise a definition for whether each of the available licenses would all abide by the same set of principles
  • However, there's no set of different licenses that are being taken and processed here, the OSI are seemingly trying to work on their own definition
• "Data Information" != "Data", and without the underlying data you can't do as much
  • This was likened to describing an Open Source project, without giving a URL - you could go and work out what the URL was, and go download the code, but in 5 years, those descriptions may lead to a 404 or a deleted repo. Or just that the data is now so out-of-date it's hard to work out what was actually true at the time
• It may be incredibly difficult (if not illegal) to create a case where the training data (in entirety) are shared - despite the fact that this mirrors what Open Source really is
  • ... and that maybe this is a good thing? It shows that you can't call it such, if you can't do that?
• Whether "open washing" is actually happening, or if companies like Meta are being careful calling it "open source" instead of "Open Source", which is OK for folks to talk about non-OSI approved licenses, so why would it be different with AI?
• That the Open Source Definition may be weakened by the vagueness of the Open Source AI Definition, opposed to the very straightforward-to-reason Open Source Definition

UTAW Panel: "I'm a tech worker, I don't need a union"

There was a good panel from the folks from UTAW around why we in tech should be unionised, regardless - or perhaps in spite of - the level of privilege we've had in tech in the past.

We heard about:

• Unions aren't just for supporting your working hours, but improving the life for everyone in communities i.e. the FLOSS community
• (Tech) Unions being able to work towards de-shittifying work, as well as the Internet as a whole, and the power they have to be able to work with governments to improve things
• Unions having made huge strides in the past at manufacturing companies, the print industry, improving gender equality, and getting us the weekend!
• A lot of folks haven't felt the need to unionise until recent years of layoffs, or replacement with AI, and hedgefunds pushing companies to outsource of offshore
• Companies working on mass redundancies/layoffs as a way to start taking back their control of our salaries - which I note on my salary page that Deliveroo had already started to reduce wages post-redundancy-announcements - and pushing wages down, which also means that we - as workers - have less control in the future, and can't as easily "jump ship"
• ... as this was often something very fortunate about the Tech world, where we could move much more easily
• One audience member noted that they've come from a teaching background, and find it "unthinkable" how we aren't unionised, and are missing out on great support and protection
• The panel noted that often people don't know they need a union until it's too late, and that we can't really expedite the process, but that people kinda have to learn that it does impact them
• That being said, we can work to talk more positively around the positive actions and impacts unions are having, and working to persuade non-unionised colleagues and friends they should do
• We heard a little from the audience around why folks weren't (yet?) unionised, leading to the fact that in some cases, they were at very small companies where i.e. there were only a couple of non-management folks, so couldn't really make an impact
• One audience member noted that "I have a smoke alarm installed in my house, well before my house is on fire" and that proactively joining a union allows them to support you when you need it, let alone working to have a recognised union that can strive to work for you
• It was noted that in the last couple of years, more folks are speaking more openly about things like layoffs/redundancies, which is great, and it's making it a little more (unfortunately) normalised, but also gives more power to us the workers to speak up
  • And again a note to talk about salaries with your colleagues!
• Discussing that being "double carded" can work i.e. if you're in a union that's recognised at your company, but they maybe don't campaign for the more systemic changes you want of the government

One of the panelists tried very hard to avoid saying "the means of production", and audibly winced when they had to say it ~19 minutes into the panel 😹

Raffle

As usual with Oggcamp, the end of the conference finished with a raffle.

I'd managed to grab a couple of sheets before the raffle started, and I'm glad I did - out of all the prizes given away, I won a Loopy Peace Student instant-print camera 👏🏼

Thanks

A huge thanks again to the organisers and the crew for putting on a great event - looking forward to the next one!