Hi John,
In the context of #1:
- I don't know if others feel the same way, but I would feel more
comfortable with a journaling file system (e.g., ext4) for important data. (I have no idea how big of a task this is)
you are certainly not alone. I'm wary to put it on the road map though.
The topic is not merely a matter of porting some file-system code but it deserves to be addressed at a larger scale, combined with a performance study. E.g., we contemplated ideas to completely separate the caching from the file systems by the means of chained VFS plugins. This work - once tackled - should also consider the concerted and graceful wind-down of the file-system stack, which remains unaddressed as of today. Such a holistic undertaking, however, is quite a big project. Too big for a side project, in my opinion.
On the other hand, I personally don't feel the pressure. The rump-kernel-based file system that I'm using day to day has not failed me yet. Let me also note that none of our current customers has recently expressed interest in funding this direction either.
And just for curiosity, can you share any info on your Sculpt UI ideas, especially "Replacing Unix/Vim-based interface of the Leitzentrale with a graphical user interface"?
Good call. Please find my collection of thoughts here:
https://genodians.org/nfeske/2020-01-06-pending-sculpt-ui
Cheers Norman