> File system check is not possible on a mounted Btrfs file system And that removes at least 50% of the reasons I'd use it. Trying to recover a partially corr

> File system check is not possible on a mounted Btrfs file system And that removes at least 50% of the reasons I'd use it. Trying to recover a partially corrupted FS is precisely when I would need a FS like BTRFS the most and is the point during which I'd expect a return of my investment. And they choose to bow out exactly then and there. Not cool. I don't want to desperately fight with Linux on a super high brightness at 80x50 of my 35" display. If that's the best your FS can offer during a time of crisis then I'm not impressed. I don't know if that's a limitation of Linux or it's a general trait of the people making such projects but I'd like to see much more investment in reliability. The file system as such is an abstraction that's not always convenient or good even. F.ex. I really can't stand installing a package and have its files splattered all over the file system! I hate that. I truly no longer care where things are, except the files I create. I however do want to be able to snapshot my entire system before installing any package and be able to roll back to it if I deem it necessary. We should focus on treating our systems like we treat GIT repos. What filesystem we use is kind of ancillary; does it help with that goal -- yes? let's try to use it then; no? then we're fine with ext4, thank you very much. If you plan to jump in and offer Nix at this point, just don't. It requires too big of an investment of time and energy. Call me when they make their CLI more intuitive. I get its idea and admire it but it seems it quickly succumbed to quirky naming that looks cute to the experts and foreign to everyone else. (We the nerds really should be restrained sometimes.)