Phillips, Addison addison at
Wed Sep 24 20:01:05 CEST 2008

> > Who cares?  It is complete bollocks in either case, as much so as
> nv-DK or tlh-AQ.
> I'm serious.
> Does anyone know which (if any) Tibetan Romanizations are in use in
> Taiwan (e.g., in teaching) and whether they differ in any material
> way from the ones already discussed here?

A Tibetan romanization used in Taiwan (or anywhere, for that matter) would not naturally use the region subtag. I mean: one can use "en-GB" anywhere, even in the United States. We don't retag the data under those circumstances. The only time this would be interesting would be if Tibetan were Romanized differently in Taiwan than anywhere else, including Tibet.

> I'm also trying to understand the generative properties of the
> proposal.
> Does the bit of information that something is based on Tongyong,
> rather than Hanyu Pinyin, come (generatlively) from the -TW-pinyin
> part of the tag or does it come (non-generatively) from the list
> "zh-Latn-TW-pinyin, zh-TW-pinyin"?

On the other hand, generative meaning has generally been problematic historically. Inferring that a collection of subtags means some unspecified variation requires tag users to intuit or have special knowledge about certain combinations. We had exactly this problem when "zh-TW" was supposed to mean "Traditional Chinese", for example.

With the pinyin case, the problem is narrower, since the subtag does encode that the content uses some form of pinyin transcription. I'd like to believe that one could just ignore the non-Hanyu Pinyin cases (especially with 'cmn' on its way), but there really is data transliterated with another Pinyin form in Mandarin. That suggests that we should provide a Hanyu Pinyin *specific* subtag and allow Tongyong et al to be registered as separate subtags. This also suggests that 'pinyin' isn't the right subtag for Hanyu Pinyin.

On the gripping hand, we could also register 'pinyin' to mean "any Pinyin". Then if somebody needs the additional distinction later (like next week), they could register subtags like 'hanyu' and 'tongyong' or, heck, '2009acad' as pinyin variations.

Again, I'd like to hear the requester's requirements clearly stated. If the requester's requirement is to tag Hanyu Pinyin to the exclusion of other forms and this list and the LSR concur with the feasibility of such a use case (which they appear to do), we should probably give them a non-inclusive subtag (even if we use my last case above to provide it).


More information about the Ietf-languages mailing list