sfscon24 🗺

SFSCON has not been on my radar. Not that I can remember at least. When I switched jobs I started looking at the open source conference offerings in Europe and stumbled over it. I visited Bolzano in February 2022. (On the trip causing these photos.) As I do I went with train from Stockholm so had to start in the early morning two days before the confernece. It was nice to arrive the evening before in Bolzano to settle in and experience a bit of this nice town in the dolomites.

The conference may be the second biggest in Europe around open source with 1000 attending. After fosdem then… It started out with a couple of single track talks in the biggest room and then it was a four track conference after that with a couple of different teams. Naturally I spent most of the time in the developers track given that is what I am. The first day - friday - is really packed with talks. And then it is a more easy going Saturday ending before lunch. In parallel there is also a hackathon. I was not up for that this time. I know it can be quite intense and coding with a laptop in a sofa is not something my body particularly appreciates these days.

The conference is free. Still lunch and coffee ws included. That was a nice surprise. I had a really nice melanzane. The venue - NOI Techpark - is a couple of kilometers southwest of the centre of Bolzano. It is possible to go there by train to Bozen Sud or by bus. I walked which took me some 40 minutes. The most direct route is through mostly industrial areas but cross the river and you can walk through residential areas instead.

A kind of badge showing sfscon stuff, my name and the mazin logo - the bird

The opening session was nice in that it set the context of the conference. Especially for a newcomer like myself.

The first real talk I attended was Open Source Businesses as an Independent, Sustainable way to Fund Open Source by Emily Omier. It was a nice walkthrough of the different options there are for making a living out of open source software.

Open source is free to use but it is not free to create

The commonly used tip jar model was not Emiliy was impressed by. Look for other business models to create a sustainable company around the open source software.

A nice introduction to AlmaLinux by benny Vasquez. I’m on Fedora as I am writing this but still dodn’t know about this new CentOS replacement.

Apisense - Easily monitor and track your REST-API data quality was a talk by two quite new developers about a tool to monitor rest APIs. I can see some use cases for this kind of software. It seemed they were mostly using it for sensors and stuff like that and that may make sense. But if you control the data flow it should be quite easy to just verify data quality in the normal code instead. I can see a use case when consuming third party APIs. Especially when it comes to anomaly detection. if an API starts to behave differently than it normally does without failing completely that would be interesting to monitor. Partial data loss for example or values out of normal bounds.

Many talks touched on the AI hype. Fulvio Mastrogianni had an interesting talk called On the ethical challenges raised by robots powered by Artificial Intelligence where he went through the current state of having robots with some kind of intelligence. He pointed out that humans normally have an observe-think-act pattern but that the current hyped AIs are not doing any thinking yet. So they can produce mechanical answers that appears to be intelligent but can not really connect the dots. So it is not likely that the LLMs will extend out in the physical world anytime soon.

A rather practical one around AI was MariaDB Vector: Why your AI data should be in an RDBMS by Kaj Arnö describing new AI features of MariaDB. Very interesting and also outlining a use case based on own data rather than the crowd sourced AIs we see mostly out in the wild.

Linters are kinda dull aren’t they. We have one now where I work for kotlin and it just pops out a bit after the fact complaining about a comma and such. Eduardo Guerra has solved it with emojis. In codEEmoji – Making code more informative with emojis he told us how. 🐘 marking too long functions, 🤷 for bad names and so on. I think I will try it. It is only for IntelliJ at the moment but may make it into VSCode as well eventually.

Then another AI talk. (To be honest I have AI fatigue.) This one quuestioning Can Test Driven Development be speeded up with Generative AI?. Jorge Melegati argued that it is possible. I’m quite sceptical about these things. If we let AI (meaning LLM) write code for us we will need to spend as much time reviwing it. I guess the future will tell if this was something to keep or not. I will continue to write my code myself in the meantime.

After this talk I went for a walk. When I came back I tried to listen to The Innovation Paradox but I don’t remember much. I think it was good stuff though.

Then onto the problem of slicing elephants with a talk called Monlith Splitter. Here we had some software that could analyze a monolith based on execution paths and git logs to come with a suggestion for how to break it up into micro services. I still don’t get how the git logs comes into the picture. Potentially you could understand what parts of the code changes at the same time or something. But to me it feels like it is quite secondary information compared with execution paths. Still an intersting problem. I am in a team that are replacing a PHP monolith. I’m not sure an approach like this would have helped us much since we (and other teams) are rewriting the thing in Kotlin and Java. In some cases we have to understand the old logic on a deeper level than just execution paths and in many other cases we can just build a better replacement without thinking too much about it. The business side of things are crucial to understand when it is important to dig into the old monolith and when we can just continue running.

And yet another AI talk called From Personalized to Programmed where Thomas Aichner put forward a few use cases where AI is not really working. No surprise there but an entertaining talk.

Ivar Grimstad of the Eclipse Foundation had a talk called Nurturing OpenJDK distribution: Eclipse Temurin Success History and plan. The main getaway for me is that Temurin is an anagram….

Then a really interesting talk about Designing open source tools for Citizen Coders. Luca Rainone rightly argued that it can be hard for non-developers to contribute. People of other professions may want to do things in an open source project but the hrudles may be too high. Making it lower would be beneficial for all. The code wil be easier to maintain and evryone can get started more easily.

The last day was a bit sparse on content and I also had a train to catch. I arrived in time for the workshop Knitting Our Internt hosted by Tommaso Marmo. It was a stroll through the history of the internet. I kinda suspected that I knew all of this already which turned out to be true. But it was inspring to experience in person nevertheless. I’m old enough to have lived through the commercialization of the internet and it is something that makes me sad. The main point from the workshop was to use yarn to demonstrate the difference between a centralized social network and a decentralized one. Whenvever there is a central node that everything goes through that node can control what content arrives with which user. It is - perhaps obivious - ironic that the internet was designed to be decentralized but is now used mostly for centralized things. I havwe been thinking about writing something about this for some time and maybe the time has come!

The last talk was an introductory to the fediverse by Tobias Diekershoff one of the developers of friendica. Nothing new here really but a good introduction and I hope some more people got inspired to make the move. And as Tobias pointed out - it is still possible to be on commercial platforms and give away data in exchange for ads.

To sum up this was a really nice confernce. Not as huge an chaotic as fosdem but leacing it with the same kind of warm feeling. This is my people and we’ll make the world a bit better each day. Actually my next conference will probably be fosdem in early February. See you there perhaps?

written by fredrik at 2024-11-09