--On Tuesday, February 13, 2018 13:50 -0500 Joe Abley
<jabley@xxxxxxxxxxx> wrote:
> On 13 Feb 2018, at 03:37, Stewart Bryant
> <stewart.bryant@xxxxxxxxx> wrote:
>
>> This may be a US vs UK English thing, but I agree with LLoyd,
>> the original seems quite correct to me.
>
> I agree. The replacement text also seems fine, though. I guess
> the new text is more correct, if we're measuring correctness
> by proportion of the audience that is happy with it.
I suggest a slightly different measure, which is tied to a long
tradition of the normative publication language of the RFC
Series being, explicitly, American English. That was, IIR, much
more strongly reflected in some of the informal documents used
after RFC 2223 and before 7322, but 7322 reflects it as a
preference for American spelling (Section 3.1) and the Chicago
Manual of Style (Section 1). The latter is definitely a
reference on American English, arguably the most extreme of them
(and, at least IMO, getting more so in recent years).
Now, while RFC 7322 is not explicit about it, although it is
implied in the spelling discussion in Section 3.1, there is also
a long tradition of allowing authors to do things in whatever
way they like as long as the language is recognizably English
and the document is consistent throughout. So, let me suggest
two theories:
(1) Following American English and the Chicago Manual of Style,
the current boilerplate construction is simply wrong, partially
because the presence or absence of negation changes nothing in
American English about subject-verb-predicate matching of
number, and the proposed new text is correct.
(2) Following the consistency rule, if the rest of an RFC is
written consistently in American English (either because the
author wrote it that way or because the Production Center
corrected it), then the boilerplate should also be consistent
with American English. Same conclusion: the proposed new text
is correct.
However, the second theory suggests a different option which, if
someone wants to advocate (and presumably do the work), I would
personally support if the RFC Editor had no objections. That
would be to allow instructions to the RFC Editor and,
presumably, a directive to XML2RFC, to specify the author's
preferred English style. If that directive specified "American"
(presumably the default, at least for historical reasons) then
the boilerplate would be shifted to match that preference. If
it specified "British", then that preference would be followed
and the boilerplate would read as Lloyd and Stewart have
suggested (with other parts of the boilerplate being made
consistent).
I do have two cautions about that plan, but they are just
cautions. First, there is a slippery slope due to the many
versions of English out there. I'm happy with the American
versus British dichotomy, but, as someone who has never been a
huge fan of the Chicago manual, I can imagine an argument for
different styles of grammatical hair-splitting based on regional
or dialect preferences. Probably we don't want to go there,
but other opinions may reasonably differ. Second, some of the
other boilerplate (not affected by the current discussion but
possibly affected if we start doing wholesale revisions to
conform to, e.g., Gowers or Fowler) could run into a problem
with assumptions that the legalese assumes US law. IANAL and
have no idea whether changes there would create any normative
problems, but have often been told that an important principle
in any litigation is not to start out by irritating the relevant
judge, including by appearing illiterate to a judge that is
sensitive to such issues. So be careful what you wish for.
best,
john