Macintosh encodings

The Macintosh encodings listed below define the upper half of eight-bit encoding vectors (i.e., positions 128..255), the lower half 0..127 being mapped to the corresponding Unicode range U+0..U+7F.

Macintosh Standard Roman

macintosh254
1
251
3 + 1
253
1 + 1
255
PDF. Refs: Guide to Macintosh Software Localization (1992) p. 263, Inside Macintosh: Text (1993).

Before Mac OS 8.5, this encoding had currency sign (¤) instead of euro (€), cf. Apple Developer Library: Text Encoding Conversion Manager Reference.

Internet Explorer substitutes the compatibility character U+2126 Ohm sign for U+0389 Omega.

Macintosh Croatian

x-mac-croatian254
1
250
3 + 2
253
1 + 1
PDF. Ref.: Inside Macintosh: Text (1993), Table A-4.

Before Mac OS 8.5, this encoding had currency sign (¤) instead of euro (€), cf. Apple Developer Library: Text Encoding Conversion Manager Reference.

Internet Explorer substitutes the compatibility character U+2126 Ohm sign for U+0389 Omega and is based on the old version with generic currency sign instead of euro.

Opera does not support this encoding.

Macintosh Romanian

x-mac-romanian254
1
246
3 + 6
253
1 + 1
PDF. Ref.: Inside Macintosh: Text (1993), Table A-5.

Before Mac OS 8.5, this encoding had currency sign (¤) instead of euro (€), cf. Apple Developer Library: Text Encoding Conversion Manager Reference.

Internet Explorer substitutes the compatibility character U+2126 Ohm sign for U+0389 Omega and is based on the old version with generic currency sign instead of euro. Furthermore, it uses Ss and Ts with cedillas as opposed to commas (added to Unicode relatively recently).

Opera does not support this encoding.

Macintosh Turkish

x-mac-turkish253
1 + 1
250
3 + 1 + 1
252
1 + 1 + 1
254
1
PDF. Refs: Guide to Macintosh Software Localization (1992) p. 275, Inside Macintosh: Text (1993), Table A-6 (neglects to mention 0xF5).

Position 0xF5 is undefined. Apple maps it to the PUA character U+F8A0 to permit ‘full round-trip fidelity’ (Apple registry of corporate-zone characters).

Internet Explorer substitutes the compatibility character U+2126 Ohm sign for U+0389 Omega.

Macintosh Icelandic

x-mac-icelandic254
1
250
3 + 2
253
1 + 1
PDF. Refs: Guide to Macintosh Software Localization (1992) p. 271, Inside Macintosh: Text (1993), Table A-7.

Before Mac OS 8.5, this encoding had currency sign (¤) instead of euro (€), cf. Apple Developer Library: Text Encoding Conversion Manager Reference.

Internet Explorer substitutes the compatibility character U+2126 Ohm sign for U+0389 Omega and is based on the old version with generic currency sign instead of euro.

Opera does not support this encoding.

Macintosh Central European

x-mac-ce254
1
252
3
253
1 + 1
255
PDF. Ref.: Guide to Macintosh Software Localization (1992) p. 265.

Macintosh Greek

x-mac-greek254
1
249
3 + 3
253
1 + 1
255
PDF. Ref. missing.

Macintosh Cyrillic

x-mac-cyrillic254
1
249
3 + 3
253
1 + 1
255
x-mac-ukrainian254
1
251
3 + 1
253
1 + 1
PDF. Ref.: Guide to Macintosh Software Localization (1992) p. 268 (likely error at position 0xB4; obvious misprint at 0xC1).

Before Mac OS 9.0, there were two versions of this encoding, both of which had currency sign (¤) instead of euro (€). The Ukrainian variant (Macintosh Ukrainian) was otherwise identical to the current version, whereas the standard variant (Macintosh Cyrillic) had cent (¢) and partial differential (∂) instead of Ukrainian G/g with upturn (Ґ/ґ). Ref. Apple Developer Library: Text Encoding Conversion Manager Reference.

Internet Explorer still uses the old versions.

Macintosh Hebrew

x-mac-hebrew250
1 + 4
203
3 + 46 + 3
244
1 + 5 + 1 + 4
PDF. Ref.: Guide to Macintosh Software Localization (1992) p. 270.

Directionality issues (Hebrew is written from right to left) are not taken into account.

Position 0x81 should normally be mapped to U+FB1F as in the Firefox implementation. Safari follows Apple’s mapping table, which uses the its decomposition instead, allegedly because some version of ‘Windows [did not] handle [U+]FB1F’. Internet Explorer incorrectly maps to U+0081 (Å).

0xC0 maps to the sequence U+05DC U+05B9, which does not appear to exist as a precomposed character in Unicode; Apple puts a PUA transcoding hint in front, which is probably better omitted in most cases. Firefox maps to U+FFFD, Internet Explorer to a PUA character.

0xC2, 0xC3, 0xC4 and 0xC5 are obsolete characters ‘used for manual positioning of nikud marks before System 7.1’, which was released in 1992; they do not appear in the Guide to Macintosh Software Localization encoding table from the same year and can probably be ignored today. Generally mapped to PUA characters with little or no font support.

0xDE has seen various mappings because U+057C qamats qatan was not originally included as a distinct character in Unicode. Maps to U+FFFD in Firefox and to a similar-looking character in Internet Explorer.

There has been some confusion regarding which of the brackets at positions 0xA8 and 0xA9 is opening and which is closing since the shape can change as a result of right-to-left processing and because of an error in an early mapping table.

Internet Explorer maps a number of bytes (including those that represent ASCII characters with right-to-left directionality) to PUA characters with questionable font support. There are also a few additional errors/deviations not mentioned here.

Support for this encoding has been removed from Firefox because of a security issue with the implementation.

Macintosh Arabic

x-mac-arabic253
1 + 1
226
3 + 26
252
1 + 1 + 1
PDF. Ref.: Guide to Macintosh Software Localization (1992) p. 264 (ellipsis apparently misplaced at 0x92 replacing í instead of at 0x93 replacing ì).

Directionality issues (Arabic is written from right to left) are not taken into account.

Position 0xBC should be expected to map to U+066D Arabic star as in Internet Explorer. Safari and Firefox both substitute U+274A (an eight-armed asterisk from Zapf Dingbats).

Internet Explorer maps bytes representing ASCII characters with right-to-left directionality to PUA characters with questionable font support.

Support for this encoding has been removed from Firefox because of a security issue with the implementation.

Macintosh Persian

x-mac-farsi253
1 + 1
252
1 + 1 + 1
PDF. The only difference from Macintosh Arabic is that this encoding uses Persian numerals.

Macintosh Thai

x-mac-thai252
1 + 2
124
2 + 129
PDF. Ref.: Guide to Macintosh Software Localization (1992) p. 274 (not including positional variants in columns 8 and 9).

Internet Explorer does support this encoding, but discards undefined bytes completely, which causes our encoding test to get out of sync.

Ad­ver­tise­ments

Contact

temp-oqgw@coq.no