Re: [apps-discuss] Review of draft-ietf-eai-rfc5337bis-dsn-1

[Dave CROCKER <[e-mail]>]

On 12/12/2010 8:40 AM, Eliot Lear wrote:
>> More importantly, the specification makes use of the =/ operator  from RFC
>> 5234, but for constructs that it does not control.  I don't know if this was
>> the original intent or not, but I am a little concerned that use of =/ in a
>> construct that is not wholly controlled by a single specification is asking
>> for confusion later.  This is an optional specification, right?
>
> After discussing this a bit with Dave Crocker, perhaps this problem is all in my
> head.  Therefore, I recommend no change to the spec on this point.


While I believe the current specification's use of =/ is not only legal but that 
it is in fact exactly the sort of use that was intended, I also note that it's a 
pretty unusual ABNF construct and that it might invite some confusion. 
Cross-referencing a rule from another specification strikes me as a big deal and 
that it's worth helping the reader as much as possible.

So I suggest having a citation for the original version of the rule be included 
in a comment associated with the new rule.

For example:

      generic-address =/ utf-8-enc-addr
                        ; updates Section 3.2.3 of [RFC3798]

d/

-- 

   Dave Crocker
   Brandenburg InternetWorking
   bbiw.net