CNS 11643 plane 1, part 1

Note: All of plane 1 has two possible EUC codes, a four-byte code prefixed with 0x8EA1 and a two-byte code without that prefix. The ICU mapping for EUC-TW-2014 treats the four-byte plane 1 codes as unassigned, while the GOV-TW CNS 11643 Word View (linked from the left column) lists only the four-byte codes. It is the two-byte codes that are listed in this chart.

Note: All four of these are underscores (horizontal versus vertical, straight versus squiggly. For some reason, older CNS charts such as ISO-IR-171 showed them aligned to the top and right of the box (possibly with the intention to use them for underlining the previous line?), as opposed to to the bottom and left as in Big5, current CNS charts, and typical rendering. This seems to have influenced UTC and Yasuoka to consider most of them unmapped, except for Yasuoka mapping the horizontal squiggly one to U+FE4B (which consequently means leaving 01-02-11 unmapped). The non-EUC ICU mapping references Yasuoka and follows suit.

Note that the "normal" underscore gets mapped to either 01-02-05 or 01-02-06. Microsoft' mapping of the horizontal, straight one here to U+2574 seems to be an arbitrary selection of a similar character not already used, and current GOV-TW mappings simply follow the Windows convention here, although CNS charts still display it as an underscore. Apple uses a PUA variation hint for a duplicate or alternate form.

Also compare 13-04-25 through 13-04-28 as used by IBM and consequently ICU, or Γ-61-30 through Γ-61-33 in IBM-950.

Note: Compare 13-01-09, Ψ-66-62.