[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [office] Fwd: New Last Call: 'Tags for Identifying Languages' toBCP
Hi David, David Faure wrote: > On Monday 13 December 2004 12:32, Michael Brauer wrote: > >>Hi David, >> >>David Faure wrote: >> >>>This might be relevant for us since we use fo:language to specify the language >>>of a run of text. Not for switching to it yet, of course, better keep following XSL >>>for now, but just in case any of you has input on the IETF draft. >>> >>>On this topic, I just noted that our fo:language is validated with [A-Za-z]{1,8} >>>(languageCode definition) >>>This basically means it's "an RFC3066 language code" but without country code. >> >>Yes. > > > So there is no way currently to tag text as "using the en-GB dictionary instead of the en-US"? There of course is a way to do that. "en-US" has to be splitted into a fo:langugage="en" and a fo:contry="GB". > How do OpenOffice users handle this problem? See above. > > >>>Shouldn't we allow things like fr_CA? (or is that fr-CA ? I'm confused by the RFC >>>talking about a hyphen, I thought it was an underscore). >> >>My understanding of RFC3066 is that it uses a hyphen. The type >>specifications for "language", "languageCode" and "countryCode" have >>been derived directly from RFC3006 and XSL-FO. >> >>RFC3066 specifies a language as >> >> >>>The syntax of this tag in ABNF [RFC 2234] is: >>> >>> Language-Tag = Primary-subtag *( "-" Subtag ) >> >>This definition is what we use for the type "language". > > OK > > >>In XSL, the datatype used for the language attribute is summarized as >> >> >>>A language-specifier in conformance with [RFC3066]. >> >>and >> >> >>>The language may be the language component of any RFC 3066 code (these >>>are derived from the ISO 639 language codes). > > OK > > >>That's from my understanding the "primary-subtag" of RFC 3006, that must >>not contain a hyphen (or underscore). > > > Why? It says "language" - which is the full thing, including a possible hyphen and subtag, isn't it? I'm not sure about this. RFC 3066 calls the full data type it specifies "language", and it does not specify a "language without a country", and no "country without a language". XSL-FO for whatever that reason does that. The XSL-FO definition in fact is not very precise, but I'm interpreting it as stated above, because - XSL-FO has different attributes for the language and the country - it talks about a "language component" and a "country component" of RFC 3066 - language components are derived from ISO 639 - This also applies (more ore less) to the "primary-tag" of RFC 3066 - contry components are derived from ISO 3166 - This also applies (more or less) to the 2nd subtag of RFC 3066. - XSL-FO also supports "xml:lang". It's type is described as "A language and/or country specifier in conformance with [RFC3066]." and it description is "The string may be any RFC 3066 code. XSL treats xml:lang as a shorthand and uses it to set the country and language properties." Michael . > -- Michael Brauer Phone: +49 40 23646 500 Technical Architect Software Engineering Fax: +49 40 23646 550 StarOffice Development Sun Microsystems GmbH Sachsenfeld 4 D-20097 Hamburg, Germany e-mail: michael.brauer@sun.com
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]