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

[Dave CROCKER <[e-mail]>]

On 12/12/2010 1:59 PM, Eliot Lear wrote:
> On 12/12/10 10:48 PM, Lyndon Nerenberg (VE6BBM/VE7TFX) wrote:
>> The base IMAP RFC (3501) made this construct easy and (I think)
>> relatively obvious.  YMMV.  If a grammar extension is globally
>> applicable to a parent specification,
...
> I take your point but then if you are calling out multiple
> specifications someone has to go on a fishing expedition to understand
> perhaps the single definition they want out of a specification.


+1

I'm sympathetic to the desire to be "efficient" in writing this stuff, but 
specifications need to be very clear and very unambiguous.  Anything that 
requires extra searching by the reader is a big negative in specification writing.

There are different way of satisfying this requirement (that is, avoiding the 
searching or confusion.)  As long as there is almost no change a reader can miss 
that it's an externally-defined rule nor miss the citation to the external 
specification for that rule, then it ought to be acceptable.

The per-rule citation approach that I've suggested is by far the safest, but yes 
there are certainly situations where other approaches will work.

d/

-- 

   Dave Crocker
   Brandenburg InternetWorking
   bbiw.net