Language subtag modification request: frr Suppres-Script Latn

John Cowan cowan at ccil.org
Fri Mar 10 14:56:14 CET 2006


Michael Everson scripsit:

> Who is "you" here? A user tagging his web page? A browser? Who 
> decides what a "normal" case is? A person? Software? Who is supposed 
> to do the omitting? A person? Software?

In each case the answer is "A user, perhaps with software assistance."

> It is interesting to note that authors of RFC306bis also had (and 
> have?) misgivings about "Suppress-Script".

It's one of those annoying but necessary compromises that make life
interesting.

> Looking at the registry, I see that "Suppress-Script" has been 
> implemented for some languages. Many other languages have no script 
> information whatsoever. I wonder how many there are of each. Not 
> being a programmer, I can't write a script to count them.  

Fortunately, no programming is required, just knowledge of the marvelous
text-processing toolkit bequeathed to us by the Mighty Ones of Bell Labs.
This is now available as a matter of course on all computers except
those running Windows, and is available as a free add-on even for those.
(Plug:  http://www.cygwin.com)

> I wonder who is working on the others

That is the collective responsibility of this list.  I will post a list
of languages in the hope that people will provide the Suppress-Script
values for them in a timely fashion.

> or why the others weren't given "Suppress-Script".

Doubtless some combination of lack of time, lack of knowledge, and lack
of certainty.

> Or of there wasn't a reason to "Suppress-Script" 
> why guidance isn't given to users with regard to scripts that might 
> be used with them. 

That information, while valuable, is not essential to the task of
constructing suitable language-tags.  Tags are to be attached to extant
documents, whose script will in general be known already.

> In "Suppress-Script", "suppress" would appear to be a verb in the 
> imperative. Who is to do the suppressing?

Once again, the user, possibly (even hopefully) with software assistance.
Here is an example.

I wish to provide a language tag for the book _Better Homes and Gardens
New Cookbook_.  RFC 3066bis says the relevant factors are the language of
the book, the script in which the language is recorded, the national or
regional orthography, and any other variant factors that are registered
or that I may wish to register.

On examination, I find that the book is written in English, in the
Latin script, and uses American orthography, lexis, and morphosyntax.
None of the registered variant subtags is applicable, and I see no reason
to erect a new distinction here by attempting to register a variant.

Consulting the syntax rules of RFC 3066bis, I construct the candidate
tag "en-Latn-US" out of the existing registered subtags.  However, I
then note that 'Latn' is the Suppress-Script of 'en', and ask myself,
"Are there special reasons to mention the script here?"  Concluding that
there are none, I tag the book "en-US".  This keeps the book's metadata
compatible with that of other documents tagged "en-US", and avoids
pointless variation between "en-US" and "en-Latn-US" that would serve
only to make the Way of the Searcher harder.

I hope this clarifies matters somewhat.

-- 
John Cowan  www.ccil.org/~cowan  www.ap.org  cowan at ccil.org
All "isms" should be "wasms".   --Abbie


More information about the Ietf-languages mailing list