Subject: Re: [email@example.com: bsdgames: number claims English output, but outputs in American]
To: Alistair Crooks <firstname.lastname@example.org>
From: Thor Lancelot Simon <email@example.com>
Date: 11/17/2003 10:10:22
On Mon, Nov 17, 2003 at 03:24:44PM +0100, Alistair Crooks wrote:
> On Mon, Nov 17, 2003 at 09:12:12AM -0500, Thor Lancelot Simon wrote:
> > It sounds like you think a witch-hunt for Anglicisms is underway. I see
> > no evidence that that is or will be the case, and do think that it is
> > both useful and necessary to pick a default dialect as a general guideline.
> I do not believe that a witch-hunt for Anglicisms is underway, in the same
> way that I do not believe there is one true spelling - I believe that there
> is room within the project for different spelling bases.
"Is room for"? Sure. However, as in any coherent (or, at least, notionally
coherent) body of written work, like an encyclopedia, newspaper, or, dare I
suggest, professionally-produced technical manual or hardware or software
documentation, I believe that, in general, a coherent style should be used;
that implies selecting a single style of spelling, grammar, and punctuation,
a task that is made far simpler by the fact that we are working in English,
where the differences in such things between the two major dialects are very
I think that viewing this issue through the narrow lens implied by such
terms as "one true spelling" is extremely unhelpful, and I don't think you'd
get far discussing the issue with an experienced technical writer or editor
in that manner, either. From my point of view the issue is simply one of
not making the reader's job needlessly difficult by using *inconsistent*
dialect where it can be avoided; so settings for one system tool should not
default to the British dialect when those for another default to the
American; when we touch up spelling and punctuation in manual pages we should
touch them up into _one dialect or the other_, not either one depending on
who is doing the touching up and his mood on a given day; and so forth.
It is an undeniable fact that the dialect used in most of the code and
documentation we have inherited from our ancestral projects is the American
dialect; thus it seems, to me, completely bizarre to set style guidelines
that would have anything else be the default. I was taught by a number of
British instructors in school and, frankly, I personally find many British
spellings far more natural -- but I find _consistency_, when we can cheaply
get it, far more important than any personal preference. And a personal
preference for _inconsistency_, which only makes things harder for the
user/reader, is something that I find almost impossible to understand.
I do have a very distinct recollection that we decided to rework the system
documentation and default-locale error messages to consistently use American
spellings, at the very least. You can find a telltale trail of commits
going back almost to the dawn of the project swapping out "ise" for "ize"
and so forth that would tend to support this.
That said, I think it's far more important to have _any_ documentation
that conforms to _any_ minimal standard of correctness and readability (in
whatever dialect) than it is to have it all conform to one set of spellings
and usages. But I do think that consistency is something we should strive
for where possible, and given that we are clearly not going to produce all
the documentation and probably not even the error messages in *both* US
and UK English, and that the vast majority of what we have is in the US
dialect, I think the shortest path towards consistency is pretty obvious.