RFC 1815 (rfc1815) - Page 2 of 6
Character Sets ISO-10646 and ISO-10646-J-1
Alternative Format: Original Text Document
RFC 1815 Character Sets ISO-10646 and ISO-10646-J-1 July 1995
That is, it's impractical to support the entirety of ISO 10646 (new
restriction or profiling can always be added), so a client needs to
know whether some restriction or profiling is being used before it
can decide whether to display the body part. Thus, it is necessary to
provide multiple charset names to each variation of ISO 10646.
For example, in Japan with Japanese windows NT, only those Han
characters already supported by MS Kanji code (mostly equivalent to
JIS X 0208 [JISX0208]) can be displayed, because no other font
pattern is commonly provided.
The other problem of ISO 10646 for Han characters is that, to display
them in quality required for daily plain text processing in
China/Japan/Korea, it is necessary to add profiling information on
which one of Chinese/Japanese/Korean the text is using. It should be
noted that this feature makes multilingual mixed
Chinese/Japanese/Korean text with ISO 10646 impractical.
Also, just as [RFC 1521] was unclear about how bi-directionality
should be supported with "ISO-8859-6" and "ISO-8859-8" which was
corrected by [RFC 1556], it is also unclear how bi-directionality
could be supported with ISO 10646. There are too much ways to
support bi- directionality. So, until some bi-directionality
mechanism(s) becomes widely supported, it is necessary to exclude
characters for languages which requires bi-directionality support
from the minimal variation. It should be noted that, though ISO
10646 is intended to be free from long term states, save for some
profiling information, introduction of bi-directionality with ISO
10646 do requires the long term states.
Combining characters also cause problems. In many countries where
combining characters based on [ISO2022] is used, there are
restrictions on how combining characters are ordered [TIS]. Without
such restriction, the result of combination is completely meaningless
which is the current state of ISO 10646. That is, if some
combination is allowed in some implementation while the other does
not support it, communication between them is difficult unless ISO
10646 is profiled to be least common set of widely supported
combinations. So, again, until combination restriction will be
developed for each language, it is necessary to exclude characters
for such languages from the minimal variation.
Conjoining characters also, may or may not be supported, which
requires another profiling.
According to those considerations, this memo defines two variations
of ISO 10646. They are "ISO-10646" as the minimal basic variation and
"ISO-10646-J-1" as the variation which could be useful in Japan.
M. Ohta Informational