[sword-devel] Correction and more :o)

[Chris Little]

On Mon, 3 Feb 2003, Daniel Glassey wrote:

> > I didn't used the same naming in both of the files (he & Hebrew).
> 
> um, question for other people. Is there any reason why we shouldn't 
> change to 3 letter ISO 639 language codes rather than the 2 letter 
> ones?

Yes, there are many good reasons to keep our current system (which is not 
quite as simple as 2 vs. 3 letter codes) rather than changing.

There is, of course, the fact that there's no good reason to "fix"  
something that is not problematic.  And I don't really predict many people 
would clamour to modify all the existing modules.

Aside from that, the current system is well thought out and follows RFC 
3066, which is also followed by XML and OSIS.  3-letter codes achieve 
little more than 2-letter codes, if one assumes to use only the ISO 639 
lists rather than the full mechanism defined by RFC 3066.  Not to 
mention... which set of 3-letter codes do you mean?  The bibliographic 
code (ISO 639-2/B) or the terminology code (ISO 639-2/T)?

That mechanism, as it is used by Sword, is as follow:
1) use a 2 letter code from ISO 639-1
2) use a 3 letter code from ISO 639-2 (it doesn't matter between B & T 
since they only differ for languages that have 2-letter codes)
3) use IANA registered language tags, prepended with i-
4) use SIL Ethnolgoue 3-letter codes, prepended with x-E-
5) make up a name, prepended with x-


Step 4 might still be unique to us.  It was chosen as the way to represent 
languages not handled by ISO 639 for OSIS, but I don't think it's in the 
schema.  But it obeys RFC 3066 anyway.

--Chris