[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

snmpconf Setcli accessor function - suggested resolution



Hi all,

Thanks to all who have participated in the setcli() discussion.
Just to demonstrate the value of such discussions, I'm afraid
that I have been swayed...

The "ugly hack" is exactly that.  I have decided that I
agree that it's not appropriate for standardization since
it is inherently a vendor-specific thing.  There have been
several strong opinions stated that seemed to push for this
way of thinking.

But we also have the reality that some vendors may need such a
thing in the short term.

My suggestion is that  rather than standardize even the name
and a set of error codes (which was all that was suggested),
we instead:

 - recognize that vendors are going to do vendor-specific
   extensions

 - document the way that vendor-specific extensions are to be
   done

 - in the section describing this, outline one possible
   vendor-specific extension that vendors COULD implement,
   including the name and a small set of return codes

This removes setcli() from any kind of standardization and
puts the responsibility for the ugly hack squarely on the
vendors's shoulders.

With kind regards,

--
David Partain                  David.Partain@ericsson.com
Ericsson Radio Systems AB      Tel:    +46 13 28 41 44
Research and Innovation        Fax:    +46 13 28 75 67
P.O. Box 1248                  http://linlab.ericsson.se/~epkpart
SE-581 12  Linköping, Sweden