John Cowan cowan at
Thu Jul 31 17:31:09 CEST 2008

Peter Constable scripsit:

> Please, let's not go overboard is trying to eliminate all redundancy
> from the descriptions. I'd rather see descriptions that have as little
> possibility of being misunderstood as possible -- and as we all know
> error-recovery requires redundancy, so I think some redundancy may
> sometimes be helpful to avoid misinterpretation.


> I'm inclined to think it a mistake for "fonipa" not to require
> "Latn".

We say that if a variant is used in a tag, then the prefix subtags SHOULD
be used as well.  The definition of SHOULD in RFC 2119 is rather strong:
"there may exist valid reasons in particular circumstances to ignore
a particular item, but the full implications must be understood and
carefully weighed before choosing a different course."

Why strongly encourage people to write en-Latn-US-fonipa instead of just
en-US-fonipa?  Since "fonipa" *implies* Latin script, there is no reason
I can see to urge people to use the "Latn" tag with it.  If they *wish*
to do so for reasons of improved fallback, that's another thing.

> Just as "Resian" qualifies "Slovenian", so also "IPA" qualifies "Latin";
> and comparably just as we would *not* consider proposing that "rozaj"
> could be used without the "sl" prefix,

Examples involving the language subtag are not really cogent, because
every (regular) tag MUST have a language subtag.

> Similarly, my initial reaction to the proposal is that having
> "zh-Latn" as the prefix in each case is fine.

I'm with Doug on this: remove "Latn" (which is not the same as
forbidding it, of course).

Using RELAX NG compact syntax to        John Cowan <cowan at>
develop schemas is one of the simple
pleasures in life....
        --Jeni Tennison                 <cowan at>

More information about the Ietf-languages mailing list