Hi Paul,
Welcome to the list and thanks for your kind words and your interest towards Genode :) From my side we can keep it public, if you don't mind, of course.
I wanted to add to my last mail for you and Mihail that our Roadmap can serve as hint in which direction we're planning to push Genode in particular, and what are therefore fields where contributing could complement our work:
https://genode.org/about/road-map
Furthermore, this year, there's one topic on the Roadmap - the FUSE/lwext4 project - that would be especially fitting for contributional work. So far, there is none of us working at it but there has been already some community effort into this direction:
https://github.com/genodelabs/genode-world/issues/193
Another idea that came up is the enhancement of the desktop tooling landscape of Genode. An appealing approach for that is indeed the development in Sculpt OS itself. This way, students could deploy software developed inside a virtual machine on-the-fly in the host system for testing it.
Regarding the time contingent, I hope you undestand that it is hard for us to estimate the duration of particular projects as it depends heavily on the experience of the developer and the approach taken. But, of course, most projects can be split up into smaller steps making the planning more flexible.
It goes without saying that each of your students that is aiming for Genode is cordially invited to seek our help through this list!
Regarding complexity and abstraction, I can only share my experience as Genode developer with you. I like especially the level of code quality in Genode that stems, IMO, from the coding style (e.g., avoid macros wherever possible), tooling (e.g., by default GCC's effective C++, GCC warnings are errors), promoted programming patterns (e.g., asynchronous programming, avoid multithreading if not really necessary, principle of least privilege), error-reducing frameworks (e.g., for type-safe sub-byte access or a very guiding framework for block-data exchange), a strict git-versioning workflow, and a broad nightly testing.
I hope this helps!?
Best regards, Martin
El 14.10.20 a las 11:40, Paul Irofti escribió:
Hi Martin,
I am the professor that recommended Genode. (Quite surprised that my student contacted you so fast and straightforward!)
Just a quick note, but maybe we can take this in private if you want, the lab work at my course involves a 5-6 weeks period where I ask my students to form small teams to tackle a large project in an open-source OS. I would love to add to my list small tasks for them to work on Genode. Of course they don't have to go into your tree :)
For kernel work I am currently suggesting they do their project on OpenBSD (just suggesting, they are free to choose) because it has not that many abstraction layers (unlike Linux) and because I am a kernel developer there since 2008 so I can help them quickly. But I would love for them to hack on Genode or Sculpt instead. From what I read so far, the code is clean and straight-forward. Of course my experience is limited here, so your word might shed some light in this regard which is why I am writing here.
Thank you for Genode and everything else!
Paul Iroft