Subject: Re: HSM deisgn goals was: RE: HSM implementation proposal
To: Ty Sarna <tsarna@endicor.com>
From: Sean Witham <Sean.Witham@asa.co.uk>
List: tech-kern
Date: 12/08/1997 18:07:12
On Fri, 5 Dec 1997, Ty Sarna wrote:

> [Try to wrap those lines at <80 columns, please Paul!]
> 
> > There are two functional issues we're dealing with:
> > 	1. Security -- we want ACL's
> 
> Well, let's say some of us want ACL's, anyway.
>

The more general goal of a more extenedable and flexable security
model is wanted by most of us. Someone else can argue how to implement
it etc.
 
> > 	2.  File assocations -- we want be be able to exexute "file.gif" and
> >	see it, or maybe "cd " into file.tar as a directory. 
> 

is it file assocation you want or support for OO features at the file
system level which can be ignored by binnaries that aren't interested.

> *Is* that what we want? Is there a reason to build it into the kernel
> instead of handling it at the application level?
> 

Depends what method is securer, cleaner and easier to manage/utilies.

> 
> 8.  Try to do things in keeping with the Unix spirit, for consistency
> if nothing else.
>

Agreed but unix has always introduced new methods for new tasks if the
old methods were not clean enough.
 
> > 	6.  VMS/NT regisrty files for both advanced file security and NDS
> >	(down the road)
> 
> In my best Andy Sipowitz voice: "Anyone dat mentions da NT registry as
> a solution for anyting, dat's gonna get yew smacked."
> 

We want the information that is held in such things as VMS/NT registry
how that information is stored and retrieved is another matter all
togther. Directory services have been goign for a long time in one
form or another: DNS, HESOID, NIS. NDS is receiving a lot of support
even from router manufactures as well as UNIX vendors so its
implecations should be assesed and support added if required.

-Sean