Re: Options

William Chops Westfield <> Fri, 30 September 1994 22:05 UTC

Received: from by IETF.CNRI.Reston.VA.US id aa09790; 30 Sep 94 18:05 EDT
Received: from CNRI.Reston.VA.US by IETF.CNRI.Reston.VA.US id aa09786; 30 Sep 94 18:05 EDT
Received: from by CNRI.Reston.VA.US id aa21302; 30 Sep 94 18:05 EDT
Received: from ( []) by (8.6.9/8.6.9M) with SMTP id QAA04420; Fri, 30 Sep 1994 16:54:37 -0500
Received: by (5.0/CRI-5.15.b.orgabbr Sdiv) id AA11680; Fri, 30 Sep 1994 16:54:34 -0500
Received: from by (5.0/CRI-5.15.b.orgabbr Sdiv) id AA11673; Fri, 30 Sep 1994 16:54:32 -0500
Received: from ( []) by (8.6.9/8.6.9M) with ESMTP id QAA04416 for <>; Fri, 30 Sep 1994 16:54:30 -0500
Received: (billw@localhost) by (8.6.8+c/CISCO.SERVER.1.1) id OAA17341; Fri, 30 Sep 1994 14:53:41 -0700
Date: Fri, 30 Sep 94 14:53:40 PDT
Sender: ietf-archive-request@IETF.CNRI.Reston.VA.US
From: William Chops Westfield <>
To: Lee Chastain <>
Subject: Re: Options
In-Reply-To: Your message of Fri, 30 Sep 94 14:39:21 MST
Message-Id: <>
Content-Length: 1164

    Yes, para 5.2 states that "implementations claiming conformance shall 
    support the following." and then ends there, implying that all subsequent
    paragraphs (such as and are mandatory.  Then in Annex A,
    A.3.5, the indication for all options is 'm'.

yes, but in para 5.1, it says:

          5.1  General Requirements

          A conforming implementation of this profile shall be
          unconditionally compliant and therefore, shall satisfy all the
          "MUST" and all the "SHOULD" requirements of the reference base
          standards and shall not implement any capability that has been
          identified by the base standards as "SHOULD NOT".Implementations
          claiming conformance to this part of DSP 2045-17506 shall support
          the following procedures as stated.

"as stated" in this case, means not supported.  Eg, it is mandatory to
support "reconnect option" as stated in rfc1123, which is to say, ignore it.
While it's rather silly to make it mandatory to ignore a particular spec,
it's less silly than requiring support for an option that has not even been
published for 20 years...