s/read/write/g
if the original file has to be world writeable because several persons
of several project groups should be allowed to modify it, creating
a new hard link into a world-open directory will allow *everybody*
to modify the original file (and not only a copy).
there are other minor differences in policy even for read-only files.
e.g. you always see the actual contents of the file and not only
a snapshot of the time when the copy was made. also if the original
file is deleted (and it's assumed that it no longer exists and doesn't
use any disk space etc.) it's not really true...
> I know of a cluster that solvs this problem by letting the users create
> their own groups: A directory is scanned once a day, and groups are
> created with their name being the file name, and the contents of the
> file being the members of the group. Almost as good as ACL's.
I know very well that having/using ACLs would fix a lot of problems
but even then I'd still consider it a bug that users are allowed
to create hard links to not-owned files as long as I can't see any
real need or benefit in this...
and right now we don't have ACLs. even if we would have them already
it's no reason not to change the link() bahaviour IMHO because it can't
fix problems with disk space accounting, quotes etc.
Harald
--
All SCSI disks will from now on ___ _____
be required to send an email notice 0--,| /OOOOOOO\
24 hours prior to complete hardware failure! <_/ / /OOOOOOOOOOO\
\ \/OOOOOOOOOOOOOOO\
\ OOOOOOOOOOOOOOOOO|//
Harald Koenig, \/\/\/\/\/\/\/\/\/
Inst.f.Theoret.Astrophysik // / \\ \
koenig@tat.physik.uni-tuebingen.de ^^^^^ ^^^^^