A Shift in GNOME’s Core Applications

For over two decades, Eye of GNOME (often shortened to EOG) was the silent workhorse of the GNOME desktop environment. It wasn’t flashy, but it did exactly what most people expected: double-click a picture, and it opened instantly. Yet, with the arrival of GNOME 45 in late 2023, a new name appeared in the lineup of “core” apps: Loupe. From that moment forward, Loupe became the official default image viewer on GNOME desktops, displacing EOG.

This decision wasn’t made lightly. GNOME has been steadily refreshing its default applications in recent years, Gedit was replaced by GNOME Text Editor, and Cheese gave way to Snapshot. Loupe is the continuation of this modernization trend. Eye of GNOME is still available in repositories for those who want it, but the GNOME team has shifted its endorsement to Loupe as the better long-term solution.

What Loupe Brings to the Table

Loupe isn’t just a reskin of EOG. It was built from scratch with today’s hardware, design standards, and security expectations in mind. At first glance, the interface looks minimal, but there’s more happening beneath the hood than many realize.

• Rust-Powered Foundation – Unlike Eye of GNOME’s decades-old C codebase, Loupe is written in Rust. This choice immediately grants it memory safety, helping avoid whole categories of crashes and vulnerabilities. For an app that regularly opens untrusted files, this is an important safeguard.

• GPU-Accelerated Image Handling – Instead of pushing all rendering to the CPU, Loupe leverages the GPU. Panning across a large image or zooming into a 50-megapixel photo feels fluid, even on high-resolution displays.

• Touch-Friendly Navigation – GNOME has been preparing for a future that includes more touch devices. Loupe fits right in, supporting pinch-to-zoom, two-finger swipes to move between images, and smooth transitions that feel natural on both touchscreens and trackpads.

• Streamlined Metadata View – Instead of burying photo information behind a separate dialog, Loupe integrates an optional sidebar. With a click, you can see dimensions, file size, EXIF data, and even location details without leaving the main view.

• Security Through Sandboxing – Image decoding is handled in isolated processes using a new backend called Glycin. If a corrupt or malicious image tries to crash the decoder, it won’t take the entire viewer down with it.

For decades, the humble terminal has been one of the most unchanging parts of the Linux desktop. Text streams flow in monochrome grids, and while the underlying libraries have evolved, the experience has remained more or less the same. Ubuntu, however, is preparing to rewrite this narrative. The distribution is adopting Ptyxis, a fresh terminal emulator designed for modern computing, and one of its standout qualities is that it leans on the GPU for rendering rather than relying solely on the CPU.

This shift is more than cosmetic. It represents a rethink of how command-line tools should perform in an era of container-heavy development, high-DPI displays, and demanding workloads. Let’s unpack what makes Ptyxis a different breed of terminal, why Ubuntu is betting on it, and what it means for everyday users and power developers alike.

The Origin Story of Ptyxis

Ptyxis is not an accidental side project. It was initially prototyped under the name GNOME Prompt by Christian Hergert, a well-known GNOME contributor also behind GNOME Builder. Early experiments showed there was space for a terminal designed from scratch with today’s GNOME ecosystem and GPU pipelines in mind.

To avoid conflicts with existing software, the project was later rebranded as Ptyxis. The application has since matured rapidly, and major distributions such as Fedora and Ubuntu have committed to it. Ubuntu introduced it in experimental form in 24.10, and by the upcoming Ubuntu 25.10 “Questing Quokka”, it is expected to replace the aging GNOME Terminal as the default choice.

A New Kind of Terminal Experience

Traditional terminals typically rely on CPU-bound rendering stacks, often through libraries like Cairo and Pango. This works fine until you throw thousands of lines of log output or try to run full-screen text-based UIs that push rendering to its limits. Ptyxis sidesteps these bottlenecks by shifting the drawing work to the graphics processor, taking advantage of Vulkan or OpenGL backends supplied by GTK4.

The result is immediately noticeable: smooth scrolling, responsive updates, and consistent performance even with massive amounts of text on screen. It’s not just about speed, either, offloading rendering to the GPU reduces CPU strain, leaving headroom for the processes you’re actually running.

Why this release cycle feels different

For most of the last decade, talk about Wayland on KDE sounded like a promise: stronger security, modern graphics, fewer legacy foot‑guns, once the pieces land. With Plasma 6, those pieces finally clicked into place. Plasma 6.1 delivered two changes that go straight to how frames hit your screen, explicit synchronization and smarter buffering, while 6.2 followed with color‑management and HDR work that makes creators and gamers care. Together, they turn “Wayland someday” into a desktop you can log into today without caveats.

The frame pipeline finally behaves

On X11/older Wayland setups, graphics drivers and compositors often assumed when work finished (“implicit sync”), which is fine until it isn’t, especially on NVIDIA, where that guesswork frequently produced flicker or glitches. Plasma 6.1’s Wayland session speaks the explicit sync protocol instead. Now the compositor and apps exchange fences that say “this frame is done,” reducing visual artifacts and making delivery predictable. If you run the proprietary NVIDIA driver, this is the change you’ve been waiting for: NVIDIA added explicit‑sync support in the 555 series, and XWayland 24.1 gained matching support so many games and legacy X11 apps benefit as well.

What you’ll notice: fewer one‑off hitches, less tearing in XWayland content, and a general sense that motion is “locked in” rather than tentative, particularly with the 555.58+ drivers.

Traditional double buffering is cruel: miss a vblank by a hair and your framerate can fall in half. KWin 6.1 added triple buffering that only kicks in when the compositor predicts a frame won’t make the next refresh, letting another frame be “in flight” without permanently increasing latency. One of KWin’s core developers outlined how it activates selectively, tries not to add avoidable lag, and works regardless of GPU vendor. It sounds simple; it feels like the end of random judder during heavy scenes.

Variable refresh is no longer a roulette wheel. KDE’s devs chased down stutter/flicker under Adaptive‑Sync, and those fixes landed in the same timeframe as Plasma 6.1. If your monitor supports FreeSync/G‑Sync Compatible and the GPU stack is sane, frame pacing is noticeably calmer.