[117] in DCNS Development
Re: Major problem with AFS
daemon@ATHENA.MIT.EDU (Ken Raeburn)
Tue Sep 24 13:03:54 1991
Date: Tue, 24 Sep 91 13:00:40 -0400
From: Ken Raeburn <Raeburn@MIT.EDU>
To: Mark Rosenstein <mar@MIT.EDU>
Cc: reidmp@Athena.mit.edu, nschmidt@Athena.mit.edu, carla@Athena.mit.edu,
In-Reply-To: Mark Rosenstein's message of Tue, 24 Sep 91 12:20:36 -0400,
Does anyone know whether Transarc has considered adding an IACL
(initial ACL, for subdirectories) to the existing ACL to address this
sort of problem?
It sounds like we need a flag per user that says whether the default
case should be readable or private--and of course, that is what the
umask is.
Is per-user the right thing, and not per-locker or per-directory? I
actually prefer the inherited ACL version; if I have a directory
that's private, or if I've granted write access to someone, I want
those attributes inherited to subdirectories, normally. It seems more
intuitive to me; my experience includes more than UNIX, however.
