Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!bloom-beacon!think!barmar
From: barmar@think.COM (Barry Margolin)
Newsgroups: comp.windows.x
Subject: Re: X over OSI
Message-ID: <21754@news.Think.COM>
Date: 6 Jun 89 02:17:25 GMT
References: <272@ucl-cs.UUCP> <8906051158.AA00566@expire.lcs.mit.edu> <26446@lll-winken.LLNL.GOV>
Sender: news@Think.COM
Reply-To: barmar@kulla.think.com.UUCP (Barry Margolin)
Organization: Thinking Machines Corporation, Cambridge MA, USA
Lines: 33

In article <26446@lll-winken.LLNL.GOV> casey@lll-crg.llnl.gov.UUCP (Casey Leedom) writes:
>  I hate to ask, but ... Is it really true that ``OSI is the only
>protocol suite a formal standard can reference''??  Have the standards
>bodies really institutionalized OSI bigotry?  (I don't know if I really
>want to hear the answer to this one ...)

The transitive closure of all standards used to implement an ANSI
standard protocol must be ANSI standards.  If an ANSI standard is
dependent upon some other standards, they must also be ANSI standards.
A protocol must be completely defined within ANSI.  The only ANSI
standard networking protocols are the OSI protocols.

This interrelationship exists in other areas.  You can't have an ANSI
standard for an X or GKS language binding for a non-ANSI-standard
language.  That's why there's currently no ANSI binding of GKS in the
C language (there's probably a draft waiting patiently for the C
standard to be ratified).

All is not lost, though.  If a protocol is reasonably independent of
the underlying protocols, then it is easy for other standards
organizations to define things in terms of the ANSI protocol.  So,
ANSI X would presumably be defined in two parts: the first part would
define the logical X protocol, and the second part would define its
implementation using OSI protocols.  DOD can then specify the TCP
replacement for the second part.  ANSI can't require a TCP/IP
implementation (because TCP/IP doesn't exist as far as ANSI is
concerned), but DOD can.

Barry Margolin
Thinking Machines Corp.

barmar@think.com
{uunet,harvard}!think!barmar