> ISO 8601 uses a hyphen character (in Unicode, U+2010 HYPHEN "‐") and minus character (in Unicode, U+2212 MINUS "−"). When using character sets which don't have these characters a hyphen-minus character (in Unicode, U+2D HYPHEN-MINUS "-") should be used instead.

This is incorrect. What ISO 8601 actually specifies is that, if the target character set is based on ISO/IEC 646 (which definitely includes Unicode), then a hyphen-minus character should be used for both cases. There is a slight ambiguity here (e.g. what if some charset contains most but not all characters in 646?) but in the case of Unicode the interpretation is clear. I believe this is a roundabout way to specify the canonical mapping for 646 and guarantee the compatibility with other charsets based on 646.