- From: John Cowan <jcowan@reutershealth.com>
- Date: Thu, 05 Apr 2001 17:15:44 -0400
- To: "I.Eck" <post@inga-eckermann.de>, www-international <www-international@w3.org>
I.Eck wrote: > I would still like to know > whether there is a tagged form of the standard codes, so that > I can import them directly in XML. I can't speak to currency codes, but trying to specify xml:lang values by enumeration is hopeless. Now that XML 1.0 2nd ed. is referencing RFC 3066, there are more than 400 different *simple* language codes, without even considering national and regional varieties, many of which are extremely important (for example, de-DE versus de-CH). It is bad practice to try to validate anything about language tags at the XML level. Although XML 1.0 1st ed. provided productions that supposedly gave the syntax, these productions were not normative for anything (neither well-formedness nor validity), and have been removed from the 2nd ed. Instead, just carry the language tag as CDATA until you get to some component which actually needs to interpret it, such as text-to-speech, spelling check, or whatever. That component will necessarily handle only a few of the thousands of possibilities. It will simply have to cope with the fact that it may receive language tags it cannot recognize, and must fall back, or give up, or both. -- There is / one art || John Cowan <jcowan@reutershealth.com> no more / no less || http://www.reutershealth.com to do / all things || http://www.ccil.org/~cowan with art- / lessness \\ -- Piet Hein
Received on Thursday, 5 April 2001 17:14:00 UTC