Inline also..
Mike Fenelon
Microsoft
> -----Original Message-----
> From: pmp-owner at pwg.org [mailto:pmp-owner at pwg.org] On Behalf Of
Wijnen,
> Bert (Bert)
> Sent: Thursday, March 10, 2005 11:53 AM
> To: 'McDonald, Ira'; 'pmp at pwg.org'; 'pwg at pwg.org'
> Subject: RE: PMP> Posted Last Call draft of Port Mon MIB (10 March
2005)
>> Inline
>>> -----Original Message-----
>> From: McDonald, Ira [mailto:imcdonald at sharplabs.com]
>> Sent: Thursday, March 10, 2005 20:19
>> To: 'Wijnen, Bert (Bert)'; McDonald, Ira; 'pmp at pwg.org';
'pwg at pwg.org'
>> Subject: RE: PMP> Posted Last Call draft of Port Mon MIB (10
>> March 2005)
>>>>>> Hi Bert,
>>>> I knew I'd forgotten to fix the four-digit years and that
>> darn REVISION clause. The uses of DisplayString were all
>> intentional for strings that can't reasonably be non-ASCII.
>>>Then it makes sense. Migth add a comment about that so that
>a novice (not as much involved) reader understands
>>> Several years ago, the SNMPv3 WG wasted a lot of time finding
>> out that using non-ASCII strings in community names mostly
>> broke existing SNMP libraries - I could certainly change
>> that one - the 'ppmPortSnmpCommunityName' object was from
>> the original Microsoft request - these printers and external
>> network adapters are all running SNMPv1 with no security,
>> but they seem to use a different read community name for
>> each printer port - we can't really throw out the Microsoft
>> requirements, because they're the motivating force for the
>> MIB.
>>>>Well, there maybe broken SNMP implementation that only accept ASCII for
>community string. But I think you are now swinging the other way to
>not accept compliant SNMP implementations that DO accept non-ascii
>charatcers in community string.
>>In any event, I find continued use of community string very scary.
>Too much opportunity for people to break in.
[MFenelon] The key here is we are not trying to use Community String as
a mechanism of security.. We know that doesn't work. It added to this
MIB because some devices use different Community Strings to segregate
request to different parts of the device.
>>> I agree with your suggestion about SnmpAdminString. I was
>> following the usage in the just-published Printer MIB v2
>> (RFC 3805) where we were told to use a dedicated TC whose
>> semantics were that a specific SNMP object contained the
>> language tag (same as here). Should we change these fields
>> to SnmpAdminString?
>>>I just posted what I noticed. I cannot mandate use of SnmpAdminString.
>As long as you do things consciously and as long as they are not
>broken, then it is you (and your group) who decides.
>>> We'll probably accept your comments as Last Call comments,
>> because we're on a short fuse to start our PWG Last Call.
>>>OK