Recent searches
Search options
Linus Torvalds[1], loosing sleep over #Linux #kernel problems since 1991. 
Today's edition:
"'"
Bah. I literally woke up and realized what the problem is. [...] I'll go back to bed, [...]
"'"[2]
Half an hour later:
"'"
> I'll go back to bed, [...]
Almost. [...] At least I understand what the problem is. Now I'm going to try to go back to sleep.
"'"[3]
[1] @torvalds
[2] https://lore.kernel.org/all/CAHk-%3Dwi6m7d-nivx10Lo%3DaGhbdk2qg-8SzjtDd9XW01LxGgAMA@mail.gmail.com/
[3] https://lore.kernel.org/all/CAHk-%3Dwi%2BWbXZcc2Sx1i-MGV2DfG4eS4Ci%2Bmrqi-PBSLSnww6qA@mail.gmail.com/
@kernellogger I feel like any creative person can sympathize with this process. The state just before sleep or immediately after awaking has a particular type of clarity that helps a mind focus on the relevant bits of a problem. :D
@kernellogger @torvalds these threads are always great for giving insight into how the kernel is developed. However, they always leave me wondering about all the hidden knowledge distributed among Linus' and the maintainers' brains, and how much of it is effectively locked down there only to be revealed from discussions like this.
@kernellogger @torvalds for example, it was cool seeing Linus go through the eventfs code, find all the code smells, and fix them with simpler code.. all because 'the VFS layer likes it this way'. However, it makes me nervous about what else does the VFS layer like. And how well-documented are these preferences and subtleties. I get that nervousness in many parts of the Linux kernel and its subsystems.
Yeah, those threads are always interesting. And yes, also, hmm, sometimes a bit worrying as one notices that a lot of information is in maintainers' brains.
But well, isn't that the case for many (all) software? Sure, sometimes it's not that bad, other time not really so. But in the end you can't force the people to write everything down. And maybe that's not worth it anyway. And maybe there are way bigger problems that should be addressed first.
IOW: It's complicated. 
@kernellogger true. I agree that this problem exists in every software ever written, some more than others.