RE: [rtcweb] Fuzzy words [was Uppercase question for RFC2119 words]

["Drage, Keith (Nokia - GB)" <keith.drage@xxxxxxxxx>]

Irrespective of the rest of this discussion, a reviewer should always address every upper and lower case usage of such verbs and ensure they are appropriate, because that is the place most mistakes are made in being absolutely precise as to meaning.

 

Keith

 

From: rtcweb [mailto:rtcweb-bounces@xxxxxxxx] On Behalf Of EXT Dave Cridland
Sent: 30 March 2016 08:25
To: Brian E Carpenter
Cc: IETF discussion list; Heather Flanagan (RFC Series Editor); rtcweb@xxxxxxxx; IESG; John C Klensin; Barry Leiba; Scott O. Bradner
Subject: Re: [rtcweb] Fuzzy words [was Uppercase question for RFC2119 words]

 

 

 

On 30 March 2016 at 00:46, Brian E Carpenter <brian.e.carpenter@xxxxxxxxx> wrote:

On 30/03/2016 05:34, John C Klensin wrote:
>
>
> --On Tuesday, March 29, 2016 08:58 +1300 Brian E Carpenter
> <brian.e.carpenter@xxxxxxxxx> wrote:
>
>> ...
>> The other words (must, shall, required, not) mean what they
>> always mean. The only argument for upper-casing them is
>> aesthetic symmetry. If a spec uses alternatives like
>> mandatory, necessary or forbidden, they are just as powerful.
>> ...
>
> Actually, when 2119 is referenced, Section 6 attaches particular
> interoperability semantics to MUST, SHALL, etc., that are not
> part of the plain-English meaning of those words.  Section 6
> seems to be ignored most of the time but cited when it supports
> an axe someone wants to grind about use of conformance language.

My claim is that even section 6 does *not* change the meanings
of the categorical words in a spec. If it says that something
must or must not happen, either the statement is redundant or
it is essential for interoperability, whether it's written
in upper case Courier New or in runes.

 

I should think you must realise that shall not always be the case.

 

But it doesn't matter. It's the SHOULDs and MAYs that require
precision in their use.

      Brian