Re: Applications Working Group

Aaron Schuman <schuman@eacces.wpd.sgi.com> Mon, 20 July 1992 19:54 UTC

Received: from NRI.NRI.Reston.Va.US by IETF.NRI.Reston.VA.US id aa17886; 20 Jul 92 15:54 EDT
Received: from wdl1.wdl.loral.com by NRI.Reston.VA.US id aa18036; 20 Jul 92 15:57 EDT
Received: by wdl1.wdl.loral.com (5.61+++/WDL-3.12) id AA08128; Mon, 20 Jul 92 12:02:44 -0700
Received: from SGI.COM by wdl1.wdl.loral.com (5.61+++/WDL-3.12) id AA08122; Mon, 20 Jul 92 12:02:41 -0700
Received: from whizzer.wpd.sgi.com by sgi.sgi.com via SMTP (911016.SGI/910110.SGI) for tsig@wdl1.wdl.loral.com id AA15776; Mon, 20 Jul 92 12:02:03 -0700
Received: from eacces.wpd.sgi.com by whizzer.wpd.sgi.com via SMTP (920330.SGI/911001.SGI) for @sgi.sgi.com:tsig@wdl1.wdl.loral.com id AA01504; Mon, 20 Jul 92 12:01:32 -0700
Received: by eacces.wpd.sgi.com (920330.SGI/900721.SGI) for @whizzer.wpd.sgi.com:tsig@wdl1.wdl.loral.com id AA28046; Mon, 20 Jul 92 12:01:52 -0700
Date: Mon, 20 Jul 92 12:01:52 -0700
From: Aaron Schuman <schuman@eacces.wpd.sgi.com>
Message-Id: <9207201901.AA28046@eacces.wpd.sgi.com>
To: tsig@wdl1.wdl.loral.com
Subject: Re: Applications Working Group
Sender: tsig-request@wdl1.wdl.loral.com

==================================================================
>>> Submissions to the tsig list: tsig@wdl1.wdl.loral.com
>>> Additions/deletions/questions: tsig-request@wdl1.wdl.loral.com
>>> Archive Server: listserv@wdl1.wdl.loral.com
==================================================================

Stan W>	I would like your feedback on this proposed group.
Stan W>	The scope of the working group could include API
Stan W>	issues for secure mail, trusted databases, secure
Stan W>	ftp and telnet, and X applications.  

Gary W>	Isn't POSIX.6 just such an API?  Won't the platform/OS
Gary W>	companies be compelled to provide .6?  I realize that
Gary W>	.6 is not done quite yet, but wonder if proposing an
Gary W>	alternative API to .6 is a wise use of TSIG time.

I appreciate Stan's predicament.  He's trying to port a
trusted multilevel distributed database to several multilevel
platforms, and their interfaces share no resemblance.
POSIX .6 doesn't begin to address interfaces to multilevel
network protocols.  If POSIX .6 were adopted and universally
implemented tomorrow, Stan would still have to deal with
syntactic and semantic differences between one company's
lrecvmsg(), another's recvlmsg(), and a third's recvmsgl().

Still, I agree with Gary.  TSIG has enough excitement
resolving technical and territorial issues with IETF.
Trying to also become the fast-path alternative to POSIX
will spread TSIG much too thin.

Most of the vendors who participate in TSIG also send
delegates to .6.  Perhaps one of those representatives
should start a trusted network API discussion.

				Aaron


.............................................................................
Aaron Schuman     415-390-1901     schuman@sgi.com     Silicon Graphics, Inc.