Variant subtag: clarification briefly sought.

John.Cowan jcowan at
Tue Oct 4 15:00:17 CEST 2005

John Clews scripsit:

> If there is a need to process text in el-Latn, which follows certain
> rules, it may not be enough.


> In passing, I haven't read the information about the "variant" subtag as
> defined in RFC 3066bis, so I would be grateful for some clarification.
> Do a notional
> el-Latn
> and
> el-Latn-[variant]
> actually differ?

el-Latn is more vague than el-Latn-[variant], which in turn is more vague
than el-Latn-[variant]-[subvariant], and so on.  Variants and subvariants
must be registered before use.  Often vague tags are good enough: the price of
infinite precision is infinite verbosity.

> What would
> el-Latn-[variant]
> add to
> el-Latn?

It specifies a particular (subnational) region, a particular
orthographic variant, a particular time period, a particular dialect
or sociolect, etc. etc.  It is used when the basic RFC 3066bis tag of
language-script-country (where "country" also includes supranational
regions) is insufficiently specific.

In the case of el, a country code is probably not useful.

> Is there a default
> el-Latn
> which therefore has certain characteristics which
> el-Latn-[variant]
> does not have?

Formally speaking, no.  In practice, probably also "no" in this case,
because there are so many formal and informal ways of writing Greek in the
Latin script, with none particularly dominant.

John Cowan  jcowan at
If I have seen farther than others, it is because I am surrounded by dwarves.
        --Murray Gell-Mann

More information about the Ietf-languages mailing list