Note that IBM ICU data for DEC-MCS is based on the glibc implementation, but this is not authoritative for this DEC charset, as the glibc "fills the holes" by mapping ISO-8859-1 characters in unassigned positions. This corresponds to a common practice, that maximizes the compatibility with ISO-8859-1 (and thus Unicode code points in range U+0000 to U+00FF).
The listed incompatibilities of DEC-MCS with ISO-8859-1 are only those 3: <!-- <OE> /xD7 <U0152> LATIN CAPITAL LIGATURE OE, instead of: <??> /xD7 <U+00D7> MULTIPLICATION SIGN --> <a u="0152" b="D7"/> <!-- <oe> /xF7 <U0153> LATIN SMALL LIGATURE OE instead of: <??> /xF7 <U+00F7> DIVISION SIGN) --> <a u="0153" b="F7"/> <!-- <Y:> /xDD <U0178> LATIN CAPITAL LETTER Y WITH DIAERESIS instead of: <??> /xDD <U+00DD> LATIN CAPITAL LETTER Y WITH ACUTE --> <a u="0178" b="DD"/> In the original DEC-MCS set, the following positions were normally not assigned (except in some extension found in more recent versions of VMS???) like they are in the iconv tables for glibc: <??> /xA0 <U+00A0 in iconv???> <??> /xA4 <U+00A4 in iconv???> <??> /xA6 <U+00A6 in iconv???> <??> /xAC <U+00AC in iconv???> <??> /xAD <U+00AD in iconv???> <??> /xAE <U+00AE in iconv???> <??> /xAF <U+00AF in iconv???> <??> /xB4 <U+00B4 in iconv???> <??> /xB8 <U+00B8 in iconv???> <??> /xBE <U+00BE in iconv???> <??> /xD0 <U+00D0 in iconv???> <??> /xDE <U+00DE in iconv???> <??> /xF0 <U+00F0 in iconv???> <??> /xFD <U+00FD in iconv???> <??> /xFE <U+00FE in iconv???> Additionally some characters were at other places (adding another incompatibility...): <??> /xA8 <U+00A4> I have also found extensions of DEC-MCS built for DBCS handling, but I don't have the details about how they work (for Japanese?) or if they used the ISO2022 encoding model, or some DEC VT controls in the encoded range /x80 to /x9F, which includes two PUA positions at /x91 and /x92... If someone has kept the technical references of those old DEC VT terminals using them, may he could answer more precisely... Ask now to Compaq or to HP?
