Re: objectIdentifierMatch on ambiguous name

["Kurt D. Zeilenga" <Kurt@OpenLDAP.org>]

At 05:04 PM 1/17/2003, Kurt D. Zeilenga wrote:
>I'd like to note, for the purposes of the WG Last Call on [Models],
>that some clarification to [Models] is needed in this area.
>        Kurt
>
>At 01:52 AM 1/7/2003, Hallvard B Furuseth wrote:
>>Kurt D. Zeilenga writes:
>>>>> Doesn't this contradict [Models, 1.3]?
>>>>I guess so, but so does your suggestion.
>>> 
>>> Yes.  So, the next question is, how do we resolve this contradiction?
>>> (I'm going to chew on this a bit before attempting to offer a
>>> suggestion.)
>>
>>I think this statement in [Models] 1.3
>>
>>>  (...) the
>>>  <descr> encoding option SHOULD be used instead of the <numericoid>
>>>  encoding option

I'm thinking that this SHOULD should be should as it appears
to be a use requirement, not an implementation requirement.

>>
>>could get the addition
>>
>>   except in contexts where the server cannot unambiguously determine
>>   which <numericoid> the <descr> represents.

I note that the original imperative applied to all
implementations (or was it uses?).

Anyways, one could generalized the addition to:
        except where use of the <descr> form is ambigious.

Which is, of course, often the case.

Hence, I suggest replacing the PARAGRAPH with:

        While the <descr> form is generally preferred, the <numericoid>
        form should be used when an unambiguous short name (descriptor)
        is not available.  See Section 6.2 for additional discussion
        of Short Names (descriptors).