month ago, I shared a project idea about windows subsystem customization, you can view it here.
here I'll share the roadmap splat to some phases:
1- phase 1: win32 familiarity:
I'll first read windows internals 7e book to get a high-level architectural overview of windows as a whole, and be familiar with win32 API and make some different and diverse application ideas that interact with different parts of windows ecosystem to have a good grasp of what I want to replace
2- phase 2: kernel mode, native applications and driver development
here I'll lower down the environment I work on to kernel mode and native applications, instead of building applications on top of win32 ecosystem and rather extend the ecosystem itself and build applications on top of my extensions, more similar to midipix project
3- phase 3: pivot to reactos and experimental PoC
I'll now work on my original idea for a custom environment subsystem but will make a simple PoC not a full blown environment subsystem, the goal here is to understand the mechanics of the tasks to make it easier to build the project on real windows NT, it's the sweep spot now to build the PoC on reactos since they're working on modern WDDM, DirectX graphics kernel and migrating to NT6.0
4- phase 4: the final task: building the actual project
here I'll start working on the actual project on real NT kernel, I'll build an installer that patches windows ISO image, build the actual environment subsystem, I'll rely a lot on reverse engineering here!
along the way, there will be a lot of theoretical knowledge I'll need, OS development theory, reverse engineering, driver development, etc...
I estimate this project to take about 4 years, and yeah it's a long period of time but this isn't an educational project I want to ship as fast as possible, it's a whole learning journey that will put me at the same tier with elite engineers :)
