Re: British or American spelling in IETF docs

--On Friday, August 14, 2026 19:51 +1200 Brian E Carpenter
<[email protected]> wrote:

> On 14-Aug-26 19:24, [email protected] wrote:
>> Exactly, enforcing one spelling will lower the work 
> 
> I don't think so. I can't write fluently in American; there are
> numerous spelling issues, and many more subtle issues. The word
> "gotten" for example really doesn't exist in British except in the
> phrase "ill-gotten gains", and in British one must write "a couple
> of bytes" whereas apparently in American it's acceptable to write
> "a couple bytes". Bottom line: I would produce text more slowly if
> asked to write in American, and it would need copy-editing anyway.

Brian,

I think your comment illustrates the problem with trying to make
specific rules about this.  In the American English I was taught, "a
couple bytes" isn't "American English", it is illiterate, just as in
British English.  Since there is no official, unified, style manual
for American English, it could also be regional variation.   For the
RFC Series, that, at least historically, illustrates the real rule,
which has traditionally been a bit more flexible, especially for
phrase constructions rather than spelling, than what several of us,
including, as Jean reported, the Style Manual, have been saying.

I hope to hear from Jean if what I'm about to say is not current
practice but, traditionally, if a construction like "a couple bytes"
showed up in an I-D the RPC or its predecessors were processing, one
of three things would happen:

 (i) They would contact the author(s) and ask if that was what
	was really intended.
 (ii) They would change it and see if the author(s) objected
	on AUTH48.
 (iii) They would not notice, or pretend to not notice, and
	just let it go under the general "author discretion"
	principle.

In either of the first two cases, there would be a discussion and
decision on a case-by-case basis, with the result often deferring to
author choice unless there are significant other stylistic issues
(and inconsistencies elsewhere in the document) with the I-D.  The
rule that has traditionally dominated all other rules has been "use
good sense", also known as "Do The Right Thing".  In the third,
unless someone complained on AUTH48, that would be the end of it.
Or, if someone did, back to the same discussion as in (i) and (ii).

And then we could move on to discussing when "bytes" or "octets" is
the correct term in that particular context.  That is a more
interesting discussion and independent of British or American
conventions :-)
 

>> and will provide consistency in documents.
>> 
>> Anyone knows if this is also the case in other SDOs, may be IEEE,
>> Broadband Forum, 3GPP, etc.?
> 
> I imagine that US-based organisations such as IEEE all use American.
> "3GPP specifications are written in (British) English."
> [https://www.3gpp.org/FTP/Information/presentations/Presentations_2
> 023/2023-drafting_rules_workshop.pptx]

More importantly, most (almost certainly not all) other SDOs use a
process in which professional SDO staff, including editors, have
control of the document text at a far earlier stage than occurs in
the IETF process.  That leads to a least two differences from the
IETF -> RFC process: (i) By the time a document approaches
publication, there are no designated/listed authors or even editors
and (ii) The style of any particular documents, and the style of all
or most documents produced by other SDOs, is whatever those staff
editors write (and therefore say it is).  There may be, and usually
are, style manuals but the interpretation of those manuals is
ultimately up to the same staff editors.  While that model has
several disadvantages relative the IETF one, it ensures a high degree
of consistency, eliminates the argument of this thread, and, fwiw,
eliminates many arguments about ownership, change control, and
copyright because the resulting documents are unambiguously works for
hire by the SDO or its parent organization.

    john