A Short Update on VSKI - Ditching Flutter
I’ve spent most of my free time past month on a passion project. Re-imagining intefaces and drafting the architecture. My first goal is to create a new UI toolkit which would make it possible to interract with AI powered systems in an effective way, but with that I don’t want to break well known user experience. A challenging task but I think it’s possible.
The first and most important phase of the project is almost complete. I am currently moving to actual prototyping and development.
At the current stage I finished gathering everything I need to create a product. Which included quite a bit of research and creating proof-of-concept code (mostly scripts).
Currently my roadmap is just a task list meant to kick-off actual feature development. My first goal is to create an application backend which will interact with os and models. The second goal is to create a simple application shell and connect it to the bakend.
Roadmap
- Initial research and PoC
- Application architecture (70%)
- Process manager
- Application bakend MVP (10%)
- Widgets/Frontend (5%)
- Application shell
Architecture
I’ve decided to use a client-server architecture for many reasons. In essence, the application consists of three parts:
- App Launcher (process manager, tray)
- Service/Daemon (backend)
- Frontend
Technologies
After playing with different GUI libraries I’ve decided to ditch Flutter it in favour of GTK4, which turned out to be a decent GUI tookit. IMHO - It deserves more attention than it gets.
Everything else stays the same:
- Vlang + C
- Tensorflow
- GTK4
Most of the development is currently in Vlang. I write about 20% of code in C when it’s more convenient, mostly due to the fact that some libs use macroses and I don’t want to rewrite them in V.
Plans
I am moving to the stage when I actually can show something except drafts. If time allows I’ll roll out first examples by the end of the March.