From: Keith McCloghrie Subject: Re: draft-ops-endpoint-mib-06.txt example syntax? To: schoenw@ibr.cs.tu-bs.de (Juergen Schoenwaelder) Date: Thu, 10 Feb 2000 17:03:29 -0800 (PST) Cc: kzm@cisco.com, tjenkins@TimeStep.com, dperkins@snmpinfo.com, mibs@ops.ietf.org, strauss@ibr.cs.tu-bs.de In-Reply-To: <200002101729.SAA17632@henkell.ibr.cs.tu-bs.de> from "Juergen Schoenwaelder" at Feb 10, 2000 06:29:12 PM Sender: owner-mibs@ops.ietf.org Precedence: bulk > You can of course write not-accessible INDEX objects during row > creation Only for those tables which support row creation. > and it makes IMHO sense to specify which subset of the > enumerated values of an INDEX object must be supported in order to be > compliant. You cannot express that formally right now if I read RFC > 2580 correctly. It does not help much that you can express refinements > for other columnar objects. Do you agree? Yes. > If so, can we also agree that the statement `must have a MAX-ACCESS > clause value of "accessible-for-notify", "read-only", "read-write", or > "read-create"' is the root of the problem? Yes. > In other words: It is OK to require that all object with MAX-ACCESS > > not-accessible are contained in at least one object group. But it is > problematic to forbid membership for an object with MAX-ACCESS < > accessible-for-notify. Almost - I agree with what you have written, but I would like to add two things: 1. it would be bad to require not-accessible objects to be present in an OBJECT-GROUP (obviously that would break current implementations). 2. it would be good if it were not purely optional for not-accessible objects to be present in an OBJECT-GROUP, i.e., to have some guidelines by which MIB writers could decide whether to include them or not. Something like include it in an OBJECT-GROUP only when there is a specific reason to do so, and otherwise advise against. The only current reason is to allow its syntax to be refiend, but other reasons might be defined in the future. Keith.