What's up for 2025
In this post I outline the plans for 2025.
Funding
With the currently accumulated funds, and current/upcoming sponsoring contracts, we can continue well into 2025. Additional sponsorships are very welcome though, so we can hire contractors, maybe organize an event, and build up some runway to get into 2026.
Priorities
The overarching priority is still to move towards a "1.0 version" after which our API can be more stable, as well as more feature-complete.
Update propagation
We already did a lot of work on update propagation and renderer / viewport refactoring in the second part of 2024. We actually hoped to finish this work in 2024, but that's how things go.
These changes will introduce backwards incompatibilities in the API of Pygfx. This means that it will cause some work and pain for downstream projects. But it will be worth it, as it makes Pygfx more future-proof, opening up more possibilities for new features and performance enhancements, with a cleaner API.
Plotting features
Another big focus will be on support for plotting features, like map projections, logarithmic ticks, axes and subplots, and more.
This will allow Fastplotlib to move to a next level, and make it easier to create other high-level scientific tools with Pygfx.
Performance
Performance optimizations continue to be an effort. Most notably to reduce the overhead when rendering many objects. Things to think about are multi-text objects, better caching to reduce overhead iterating over all objects, and making use of WebGPU's render bundles.
Remote rendering
We also have plans to improve support for remote rendering to rendercanvas.
This includes research to optimising performance of sending the rendered images
to another place, which will also benefit jupyter_rfb.