Please delete the subject page "MediaWiki:Privacy/zh-tw"

Not completely.

• "zh-cn" represents a part of of "zh-hans", but it actually means "cmn-hans-cn". "zh-hans" also covers "cmn-hans-sg"
• "zh-tw" represents a part of of "zh-hant", but it actually means "cmn-hant-tw": "zh-hant" also covers "cmn-hant-mo" for example.
• "zh-hk" represents a part of of "zh-hant", but it actually means "cmn-hant-hk" (but it is also mixed terminologically with Cantonese/Yue "yue" and terms borrowed/adapted from English).

The mapping from "zh" to "cmn" comes from the registration in the IANA database for BCP47, where Mandarin is the default language within Chinese languages "zh", which also includes some other Sinitic languages like Hakka, Minnan, Cantonese/Yue or Wuu... but not all, for example Tajik [tg] is not part of the "zh" macrolanguage, even if it is Sinitic).

Chinese "zh" has many different "dialects" (actually different languages, that's why it is a "macrolanguage") which are widely different when voiced (so much that they are not easily mutually intelligible when spoken), but using two shared main orthographies ("hans" and "hant"), with only some local differences (e.g. in Hong Kong), but also additional non-Han scripts (notably Latin for the Minnan orthography, or for Pinyin transcriptions of Mandarin which is also a common alternative used by IMEs for "hans" and "hant"; with the Latin script there are also other romanization systems than Pinyin, such as Wade-Giles mostly used by Chinese speakers in US). With their most common written form using the Han script, all these languages become somewhat intelligible with each other between China Taiwan, Singapore, or other Chinese writers/readers in other countries.