Path: utzoo!utgpu!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!mailrus!ames!xanth!mcnc!decvax!decwrl!sun!pitstop!sundc!seismo!uunet!steinmetz!davidsen
From: davidsen@steinmetz.ge.com (William E. Davidsen Jr)
Newsgroups: comp.arch
Subject: Re: Operating Systems (Re: archimedes)
Message-ID: <12750@steinmetz.ge.com>
Date: 6 Dec 88 21:08:02 GMT
References: <12633@steinmetz.ge.com> <6191@killer.DALLAS.TX.US> <12655@steinmetz.ge.com> <343@maxim.ERBE.SE>
Reply-To: davidsen@crdos1.UUCP (bill davidsen)
Organization: General Electric CRD, Schenectady, NY
Lines: 23


  I once did an o/s design as a consultant (which was never implemented)
and one of the things in it was the concept of a "per file" permissions
manager. If there was no permissions manager access worked a lot like
UNIX, but with a permissions manager when an access was made the info
was passed to the appropriate entry in the permissions manager. The
entries were (not original here) open, close, read, write, seek, and
tell.

  For example, if you wanted access control lists, you could write a
permissions manager which checked some ACL before allowing access to a
file. How to do it? If you like a header on the file you can do it, just
offset all accesses. Do you like a shadow file giving the ACL? You could
do that, too.

  I intended that anyone could have any kind of access control scheme
desired, with anything layered over it, even the 29 flavors of file from
VMS. After the design was pretty well along it because obvious that the
project was too big to continue, and was dropped.
-- 
	bill davidsen		(wedu@ge-crd.arpa)
  {uunet | philabs}!steinmetz!crdos1!davidsen
"Stupidity, like virtue, is its own reward" -me