[Gnu-arch-users] Re: extended attributes

[Tom Lord]

    > From: Aaron Bentley <address@hidden>

    > Perhaps a fully transactional filesystem was never what was needed. 
    > Maybe all that's needed is a variant of rename(2) that only changes the
    > number of links, atime and mtime.

Huh?  Can you explain in more detail what you are thinking of because
I don't get it from that description.   mtime?   really?  What?!?

    > > It's not very surprising -- the semantics come out of an
    > > implementation process which amounts to "Ok, like multics, except with
    > > a relentless focus on `what is that we can quickly (and actually)
    > > implement?'."

    > Well, yes.  But some things that could be quickly implemented were not
    > because the need wasn't apparent.

Sure.  Informally, it feels like the original idea was "every file is
a sequence of bytes; directories being files that can only be written
to via dir-modifying system calls" and then it's all downhill from
there.  (You know, extended attributes ain't nothing -- you can
probably still find people with principled objections to symlinks.)

-t