Re: Question about BCP 14 / RFC 8174

["Rob Wilton (rwilton)" <[e-mail]>]

I think that writing “MUST xxx unless yyy” is more precise and clear than” SHOULD xxx unless yyy”.

Picking up on Brian’s prose, arguably writing “MUST xxx UNLESS yyy”, “MUST NOT aaa UNLESS bbb” would be even clearer to readers.  A slightly out-there suggestion could be to update RFC 2119 to remove SHOULD/SHOULD NOT and introduce MUST … UNLESS and MUST NOT … UNLESS as their replacements.

That would make specifications more precise and prevent folks from using SHOULD to sit on the fence.

Kind regards,
Rob



From: John C Klensin <[e-mail]>
Date: Friday, 29 August 2025 at 03:49
To: Brian E Carpenter <[e-mail]>, Toerless Eckert <[e-mail]>
Cc: [e-mail] <[e-mail]>
Subject: Re: Question about BCP 14 / RFC 8174


--On Friday, August 29, 2025 11:45 +1200 Brian E Carpenter
<[e-mail]> wrote:

> Toerless, I agree with you but again, my concern is not really
> about MUST/MANDATORY or MAY/OPTIONAL. It's about
> SHOULD/RECOMMENDED, where it might often be better to write MUST...
> UNLESS without BCP 14 really being needed.
>
> As in "You SHOULD look both ways before crossing the road" vs "You
> MUST look both ways before crossing the road UNLESS you are being
> chased by a grizzly bear." That's the BCP 14 interpretation of
> SHOULD, but we have often seen evidence of developers interpreting
> it as "You don't really have to look both ways before crossing the
> road."

And, although this has been said many times before, if we more often
insisted on "SHOULD ... unless" discussions where they are possible
-- e.g., "You SHOULD look both ways before crossing the road unless
you are being chased by a dangerous animal" -- we'd probably have
fewer misunderstandings of what SHOULD implies, in particular the
misunderstanding about it having a definition closer to "we'd prefer
that you do it that way, but it is really up to you".

And I agree with Toerless, and apparently you, about the BCP 14
language.  I'd go a tad further and suggest it is better adapted to
Technical Specifications than to Applicability Statements, better for
Applicability Statements than to BCPs, and better for BCPs than
Experimental or Informational documents.  Maybe eventually a question
for offspring-of-PROCON.

> I think operators are much less likely to make that mistake,
> whether the RFC cites BCP 14 or not.

Probably.

best,
  john