Presenting FEX, a translation layer to run x86 apps and games on ARM devices: Learn why x86 is such a pain to emulate, what tricks and techniques make your games fly with minimal translation overhead, and how we are seamless enough that you'll forget what CPU you're using in the first place!
ARM-powered hardware in laptops promises longer battery life at the same compute performance as before, but a translation layer like FEX is needed to run existing x86 software. We'll look at the technical challenges involved in making this possible: designing a high-performance binary recompiler, translating Linux system calls across architectures, and forwarding library calls to their ARM counterparts.
Gaming in particular poses extreme demands on FEX and raises further questions: How do we enable GPU acceleration in an emulated environment? How can we integrate Wine to run Windows games on Linux ARM? Why is Steam itself the ultimate boss battle for x86 emulation? And why in the world do we care more about page sizes than German standardization institutes?
This talk will be accessible to a technical audience and gaming enthusiasts alike. However, be prepared to learn cursed knowledge you won't be able to forget!
Flatpak has become the de facto standard for distributing desktop Linux applications across distributions, and the XDG Desktop Portal ecosystem continues to evolve alongside it. This talk provides a status update on both projects: what's shipped recently, what's in progress, and where things are headed. We'll cover new, updated and planned portal interfaces, ongoing efforts to modernize the internals of Flatpak and preparations for a next-gen Flatpak. Whether you're an application developer, a desktop environment maintainer, or simply a curious user, come find out where things stand and how you can get involved.
Over the past year Modal Collective has been working towards bringing local-first sync to native GNOME apps. As part of this, we built Reflection, a collaborative GTK notes app, improved p2panda APIs and documentation, organized developer events, and did user testing.
Another important part of this project was thinking about new user interfaces paradigms. Building complex local-first apps requires answering a lot of new, difficult questions in order to do seemingly simple things. How do you save and and delete documents? What does it mean to have a user account? How do you manage groups and permissions? What novel threats do we need to consider?
In this talk I'll discuss some of the design work we did as part of the Reflection project. This includes user interface experiments that are not implemented yet, questions we don't have clear answers to at the moment, and what we think needs solving most urgently.
In 2024 the KDE community elected "Streamlined Application Development Experience" as one of their goals for the coming years.
In this talk we are going to explore what this means, what has been achieved since then, and what the future of developing KDE applications will hold.
The EU is moving off proprietary platforms and onto Linux in the name of digital sovereignty. That gives us a unique opportunity to build up the Linux app ecosystem. Specifically, office suites are going to be singled out first, whether you're a government, a company, or a non-profit working with EU offices.
But the many open-source document suites we have today — LibreOffice, Collabora, ONLYOFFICE, and more — most still have many gaps that keep the experience from feeling really complete: low fidelity between editors, broken formatting, dated UI, poor use of AI, and real-time collaboration that feels bolted on rather than built in.
Using ONLYOFFICE as a case study, I'll walk through three things: the engineering decisions we made to address these gaps, why those decisions made Linux users, and not Windows or Mac, our largest base, and what the rest of the Linux app ecosystem can take from any of it. If you are involved in the building of any kind of Linux app trying to win a serious user base, you are in the right room.
We are living through an unprecedented boom in indie app development on our platforms, which introduces new challenges at a larger scale.
Every week, Flathub gets flooded with new, high-quality and innovative projects, but how many survive the test of time? Is there life after EOL? Can AI help?
Pressing the fork button is the easy part. Find out the social rules around forking, how to lay the foundations for long-term sustainability without burning out maintainers.
TuxDoctor: A Hardware Diagnostic Tool for Mobile Linux
Mobile Linux devices span a wide range of hardware, distributions, and UI stacks making it hard to know whether a phone or tablet is actually
working correctly. TuxDoctor is an open-source diagnostic tool written in Rust that brings a consistent, reproducible way to test mobile Linux hardware.
It runs a suite of tests covering connectivity (Wi-Fi, Bluetooth, cellular, NFC), sensors (accelerometer, gyroscope, magnetometer, light,
proximity), audio and haptics, camera, screen, biometrics, and battery health. Each test reports a clear pass/fail/skip status and produces a
shareable local report including full device and OS information.
TuxDoctor ships a GTK interface alongside a CLI and a background daemon, and is packaged for both postmarketOS (Alpine) and Mobian
(Debian). This talk will cover the architecture, the challenges of writing portable hardware tests across diverse ARM devices, and how the tool
can help users, developers, and porters evaluate device health.
https://tuxdoctor.com/
The Linux application ecosystem continues to expand as new hardware platforms and distribution channels bring Linux to more developers and users worldwide. One of the most promising developments in this space is the rapid growth of the RISC-V ecosystem, which introduces an open instruction set architecture designed to encourage innovation and collaboration across hardware and software communities.
With the introduction of the RVA23 profile, RISC-V platforms are becoming more standardized and software-friendly, enabling better compatibility for operating systems, development tools, and applications. These improvements help reduce fragmentation and make it easier for developers to build and distribute applications that run reliably across multiple RISC-V devices.
This session explores how emerging RISC-V platforms—particularly new developer systems and laptops—are helping expand the reach of Linux applications and create new opportunities for developers. We will discuss how standardized profiles like RVA23 support better software portability, how open hardware platforms can attract new contributors, and why growing the developer base is essential for sustaining the Linux app ecosystem.
Attendees will gain insight into how hardware innovation and open-source collaboration can reinforce each other, enabling a broader and more resilient Linux ecosystem. The session will also highlight practical ways developers, distributions, and communities can participate in building a thriving Linux application landscape for the next generation of computing platforms.
Spectrum is a desktop Linux operating system with a focus on security. Each application instance, and some drivers, are run in their own virtual machines, but this virtualization is designed to be as transparent as possible to users using standard mechanisms like cross-domain virtio-gpu and XDG Desktop Portals in a new way.
Other desktop Linux systems struggle to take full advantage of the security architecture of Wayland, Flatpak, etc. due to their need to maintain compatibility with applications that were not designed with modern interfaces in mind. This talk will demonstrate how Spectrum is able to use these technologies to their full potential thanks to its positioning, which lets us prioritize security over legacy features, while still aiming for maximal compatibility with the modern Linux application ecosystem and user expectations.
This talk is a continuation (and partial rehash) of last year's "TinySPARQL, LocalSearch, and the future of search in GNOME", with the additional knowledge gained over the course of a year.
If you want to hear about fancy concepts such as embeddings, vector databases and similarity search. This might be your talk.
In recent months I've been exploring further the adoption of BuildStream in KDE with the goal to improve how we deliver our software. I'll discuss how this affects the compilation of KDE's flatpak runtime, KDE's flatpak applications and KDE Linux.
This work will be contra-posed to the status quo, namely the usage of flatpak-builder, ArchLinux packages for KDE Linux and the many other tools we might find on the way.
The bulk of the work in this presentation has been sponsored by Codethink.
An exploration of the mistakes we made the first time around, how the world around us has changed, and how the next generation of Flatpak and Portals will look like.
We'll look at KDE Plasma's magnificent crash reporting system and how it enables users to submit crash reports conveniently. Through the amount of aggregated crashes, developers can prioritize issues correctly and get enough data points to diagnose and resolve the underlying causes.
From collecting cores, over preparing the submission, to eventual presentation to the developer, we will look at the entire journey a crash takes in KDE Plasma.
Emergence is (primarily) a data replication/synchronization mechanism for TinySparql databases, based on W3C standards, and oriented to personal data across multiple personal devices.
A technical overview was presented for this project at Guadec 2023 in Riga, mostly as a research concept. This talk will cover the journey in materializing the project, the learnings that can be extrapolated to other user data, and some musings on how a more englobing mechanism ready to use in/across desktop environments could look like.
Join a founder of LibreOffice for an in-depth look at the current state
of open source Office applications on the Linux desktop. This talk will
explore the evolving landscape of productivity tools, comparing
different approaches to native platform integration and the ongoing
challenges of keeping pace with rapidly changing desktop toolkits.
We will examine modern packaging and distribution strategies, including
native packages, Flatpak, and Snap, and discuss the practical
implications for users and administrators. Topics such as font
management, printing, and the balance between online and offline
collaboration will be covered from a real-world deployment perspective.
A key focus of the session is the growing importance of digital
sovereignty. Office suites are not just productivity tools - they are
central to how organizations create, store, and control their data.
Using real migration efforts such as those in Schleswig-Holstein and the
Austrian military, we will explore how public institutions are reducing
dependency on proprietary vendors by adopting open standards like ODF,
and open source solutions. This includes regaining control over document
formats, update cycles, and data flows, as well as the broader
challenges that come with such transitions.
Finally, the talk will showcase recent design improvements and new
features in Collabora Office on the desktop, highlighting how modern
open source solutions continue to evolve to meet both user expectations
and strategic requirements.
Whether you are a developer, system integrator, or FOSS end-user, this
session provides insight into the technical and organizational realities
shaping the future of Office on the Linux desktop.
Project sovereignty isn't just about hosting infrastructure, it's about shipping an experience as intended by the people who create it.
In this talk I will explain why I think GNOME and KDE should not only make their own operating systems, but to use them in a way to drive modern linux development - all for a better user experience. I will cover how this matters for each pillar of discussion:
Digital Sovereignty - More than ever projects are working to ensure supply chain security as well as infrastructure independence. I will talk about how open standards make that happen. We overdiscuss implementation details sometime, I want to focus on the "APIs" we can use to make this happen.
AI and Machine Learning - I will discuss the impact of AI in Linux infrastructure and how it will affect our projects and communities.
Ecosystem Growth, Innovation, and Platform diversity - I will discuss the importance of growing a developer community around modern Linux.
And where do GNOME and KDE place their operating systems? Linux distributions have been strategic partners for these organizations for decades. Who is competing with who? I plan on discussing how the cloud native community handles this sort of situation in a way that gets everybody wins.
The work done by the UAPI Group and cloud-native projects like bootc have certainly caught the eyes of infrastructure folks. I will explain how we can leverage this massive investment in modern linux for the desktop. I have been collecting 5 years of data on this, and look forward to sharing it with the community. There will be charts and dinosaurs in this talk.