HP 9000 User Manual page 151

A
HP15 definition use a set of bytes with the high bit set to differentiate between
8 and 16 bit data. If the high order bit is zero, then the byte represents a
one-byte ASCII character.
Otherwise, the current byte may represent the first byte of a two-byte character
or a one-byte non-ASCII character (correct interpretation is provided through
system tables). Since the high order bit acts as a flag bit, only 15 bits remain
for data, hence, the term HPI5. While this representation allows 8-bit data to
be distinguished from multi-byte data, there is one restriction. A byte stream
must be examined sequentially. For an "arbitrary" byte it is not possible to tell
if the byte represents a single byte-character or the second half of a two-byte
character.
In addition to HPI5, NLS defines support for EUC (for Extended UNIX Code),
an encoding scheme defined by AT&T UNIX Pacific, Ltd. The EUC encoding
scheme provides a general template for defining codesets and interpreting a
byte stream. EUC allows four distinct code sets (CSO through CS3) to co-exist
in a byte stream by restricting the legal bit patterns for each set to a specific
template. Under the EUC encoding scheme ASCII characters are uniquely
identified by a most significant bit of "0" and occupy CSO; the EUC template
restricts the most significant bit of
all
non-ASCII characters to a "1", even the
second byte of two-byte characters.
Currently HP- UX NLS supports japanese.euc (U JIS), a codeset that supports
the Japanese language. This support includes the
2
byte codesets: CSl, and
CS2.
A-4 Special Topics for HP's 16-bit Interfaces