The IETF "Language Tag Registry Update" working group has noted that lots of data is tagged as "zh-Hant", regardless of whether or not it is pronounced as Cantonese or Manderin.  For video and audio however, this doesn't allow a fine enough distinction, and so the LTRU is working on revising RFC 4646/4647 and the registry to allow for new tags to distiguish Cantonese and Manderin from the "macrolanguage" of Chinese.

So in the future we should expect to see "cmn" and "yue" tags instead of zh.  The LTRU is still a bit in flux about the details, but it is clear that in the future newly tagged data will use "cmn" and "yue".  This is going to cause "an interesting time" since lots of legacy data, resources and systems will continue to use the zh tags.

User configurations may need to change, such as allowing both "cmn and zh" in a web browser's language configuration.  Applications and systems may also need to change to provide "cmn" resources if "zh" was asked for, or vice versa.  Content providers may also need to retag existing data to distinguish between Cantonese and Manderin.

With these types of changes, the adoption rate is usually quite varied, so expect some applications and content to shift rapidly to using the new recommended names once that new standard is created.  Other data and systems will probably remain unchanged for a very long time, leading to very interesting scenarios when those environments communicate with each other.