Hi Joan,

I like starting at zero as well.  It does present some issues
for SNMPv1 implementations that might want to support the MIB.

Even though v1 has been declared historic I would tend to wager
that it is still the norm in the industry even though the IETF
might wish that it were not.

So there is a choice between new and elegant, and old and more
widely adoptable.  I could go with a majority vote on this one,
or just a plain old good thrashing by Mr. Rose or one of his
peers.  Either way I am sure that it will be fun.

The MIB is in the security/syslog working group and there might
be an -02 version making the rounds.

Thank you,

Bruno



-----Original Message-----
From: Joan Cucchiara [mailto:[EMAIL PROTECTED]]
Sent: Monday, August 26, 2002 6:20 PM
To: Pape, Bruno
Subject: [Fwd: Re: draft-ietf-syslog-device-mib-00.txt - Enums starting at
zero]


Hello Bruno,

I posted this to the mibs email list (not sure if you
saw) but would like to request that this
enum start at zero.  Could you put this in the next
revision of the draft?

Also, which working group is this MIB being developed in?
And when can we expect to see a new rev of this MIB?

  Thanks, Joan

-------- Original Message --------
Subject: Re: draft-ietf-syslog-device-mib-00.txt - Enums starting at
zero
Date: Mon, 26 Aug 2002 14:40:01 -0700 (PDT)
From: "C. M. Heard" <[EMAIL PROTECTED]>
To: [EMAIL PROTECTED]

On Mon, 26 Aug 2002, Joan Cucchiara wrote:
> My question is about what the current practice is
> for starting enums at zero when there is a reason
> for starting at zero.

Values of enums in SMIv2 are not restricted like they
were in SMIv1 -- they can have any integer value.
There is no longer any reason to require them to
start at 1.

> Specifically in this MIB, there is an enum:
>
> SyslogSeverity  ::=  TEXTUAL-CONVENTION
>     STATUS  current
>     DESCRIPTION
>         "This textual convention maps out to the severity levels
>          of syslog messages.  The syslog protocol uses the values
>          0 (emergency), to 7 (debug)."
>     SYNTAX  INTEGER {
>                       emergency(1),
>                       alert(2),
>                       critical(3),
>                       error(4),
>                       warning(5),
>                       notice(6),
>                       info(7),
>                       debug(8)
>                     }

Re-numbering the enums from 0 to 7 to match RFC 3164 would make
more sense than what is done above.

> At one point having enums start at zero in SMIv2 was
> not encouraged due to backwards compatibility with SMIv1,
> but am wondering if this is still the case?

No, not any more.  We keep SMIv1 around because we don't
want to have to re-write perfectly good SMIv1 MIBs, but
essentially all new work these days is done in SMIv2,
both within and outside the IETF.

//cmh

Reply via email to