Broome, Karen Karen_Broome at
Tue Aug 5 00:09:21 CEST 2008

Addison writes:

>Note that 'cmn' would be added as a prefix when available.

And zh-cmn too?

This is where I think we create problems by not having a single preferred tag for each of the Chinese variants. We know in the future "zh" will not be a preferred tag, so I don't think "zh-pinyin" should be allowed today. I wonder if this tag is being registered now because the registrant intends to ignore the preferred Chinese tagging chosen by the committee in the future and knows that in the RFC 4646bis era, it's less likely that new "zh" tag variants would be allowed.

I strongly believe that this request should be associated with the "cmn" or "zh-cmn" tag ... but three valid prefixes for the same thing? Might as well throw zh-guoyu in there too. If you don't, the treatment of these tags is inconsistent. This is where I start to see this standard as something less desirable than its predecessor for most uses of this work -- it allows too many options for the same thing and the exceptions are becoming harder and harder to explain. We can see the exponential nature of these options in requests like this.

For the purposes of registration, I think we should consider "zh" deprecated. It may be a popular tag with a lot of legacy classifications, but that doesn't make it specific enough to span use across the written and spoken language identification needs of the Internet today. "Zh" is a Bad Tag and we should discourage its use moving forward. "zh-cmn" is the preferred tag for Mandarin today. If "pinyin" is a Mandarin-only variant, it should use the Mandarin tag.


Karen Broome

More information about the Ietf-languages mailing list