Sorry for the delay in following up with this discussion.
I thought about this and I think the best solution is to add an
extension to FOP to handle the cases that cannot nicely fit in language
plus country pattern. FOP already has extensions (new XML elements or
attributes) to handle situations that do not fall inside the scope of
the XSL spec, and this will be another situation. In practical terms, I
will remove from the repository the files that were incorrectly named
and that triggered this discussion. FOP users that want, say, to
hyphenate classical Latin, will then need to use the original file name
in the repository.
I will also use this opportunity to clarify that most FOP users use it
to generate PDF and other print formats. I think that among those users
some do indeed use Latin!
On 3/11/16 11:26 AM, Mojca Miklavec wrote:
On 11 March 2016 at 10:59, Arthur Reutenauer wrote:
Yes, of course, BCP 47 is a standard of the IETF, the Internet
Engineering Task Force. But that's not directly relevant, what's more
important is how XSL tags languages, and it seems it only supports ISO 639,
with country and scripts in a different field (following ISO 3166-1 and
ISO 15925 respectively); even for languages it only supports ISO 639-1
and 639-2, not -3: https://www.w3.org/TR/xsl/#language
Thanks for the link. I'm sorry for not properly understanding the scope of use.
So in this particular case there is probably nothing we or OFFO
developers can do to support two sets for one language unless there
are other private fields.
(And 'mul' is explicitly turned into undefined, so mul-ethi has to be changed.)
Nevertheless this doesn't mean that we couldn't try to get an official
tag registered for Latin.
Mojca
