[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: Get rid of "suggested minimum upper bound"



Kurt D. Zeilenga wrote:
At 08:11 AM 2002-03-11, Michael Ströder wrote:

Kurt D. Zeilenga wrote:

I don't think we can just replace noidlen with numericoid
in the AttributeTypeDescription ABNF.  That would likely cause
interoperability problems.
I agree clarification is needed.  Minimally, the words
"maximum length" in 4.2 needs to be replaced with
"suggested minimum upper bound".

Maybe what we should clarify first whether there's any
>>known implementation which believes to do anything useful with it.

Lots of defined schema, using this ABNF, contains suggested minimum upper bounds. A large number of client and server implementations support updating this schema via the protocol. Changing the ABNF would require all these implementations refuse to accept this defined schema. That would be bad thing.

Kurt, you are already one step further with thinking about how to tweak the standard documents.


Basically I asked for feedback of implementors. If anyone relies on the "suggested minimum upper bound" (formerly known as "maximum length" ;-) and how implementations are interpreting the meaning of it.

Afterwards (or parallel) we can discuss the ABNF off course.

Ciao, Michael.