International Unix

[Erik E. &]

The point about the kernel not having much ASCII dependent code is well
taken, however, I was thinking (and expounding upon) the whole of UNIX,
which most specifically includes the entire lot of ASCII ridden utility
programs. For an inaccurate survey (it will grossly underestimate the
number of programs that will have to be changed), do something like
this:

	cd /usr/src
	egrep -l 'include.*ctype.h' *[ch] */*[ch] | wc -l

This will give you the number of files that include `ctype.h' which
is explicitly ASCII dependent.

Someone else said that different European languages (we're ignoring the
orient for the moment) use the same glyph or letter in different order,
or with completely different meaning, and therefore an international
character set can't be done.

If that's the case, then each country will end up writing its own
version of UNIX based on the national character set (like the French
have done, and the Japanese are doing now).

The real goal we're shooting for is the international exchange of
information. There's nothing stopping the Europeans or the Japanese
from building their own computers on incompatible character sets. And
they do so. However, it is hard, slow and tedious to translate data
from a Japanese computer in (say) kanji, to English/ASCII.

I think that we're attacking the wrong problem, however. Instead of
attempting the technological solution by teaching computers some large
`n' number of languages, we should attack the basal cultural problem by
developing and widely teaching a common intermediate language.
Esperanto, anyone?

	Erik E. Fair	ucbvax!fair	fair at ucbarpa.BERKELEY.EDU