The algorithm exists in my ezrsnapshots bash program (which I probably
should just make a page for)
That would be nice.
There's a strong point raised that I very much get, that mucking with a
stable and an established code and user behavior for something like this
causes hesitation for good reason. It's easy to play with ideas in alpha
code when nobody is using it yet.
But we should also ask ourselves if it's even a good idea to even
place in alpha code.
The time and date does exist. It's in the time stamp. That's been
said over and over and over.
yes but it makes sense in the context of a code that keeps fixed file names anyway, which just happens to be extremely useful for simplifying everything else.
Creating more code to solve a problem that doesn't exist isn't a good thing.
I agree as far as filenaming for its own sake. ls -tral does the same thing.
BTRFS (a COW snapshotting filesystem) has an fsck now, so all of this stuff
might be silly soon anyway.
It's also still not ready as a production filesystem used for backups, IMO.
(and also that of the developer, Tso.)
I also agree, which is why I said soon maybe. The fsck seemed rushed out to list one point.