It is worth stating that ICS-2 could probably work for 99.9% of applications
but not allowing  some characters is  unacceptable as it doesn't allow some
word s , also since  a new popular word may be introduced as a new composite
character it is a (slowly) changing character set just like our dictionary
changes though in the majority of cases new words use 2-4 existing
characters) .

 

.NET is not illegal because  UCS-2 is only internal, the output of .NET is
nearly always ascii, UTF-8 or UTF-16.  I was merely conveying that UCS-2 is
a legacy standard since it can't represent all reasonably used characters
in China ( and Japan , Hong Kong and  Taiwan) and is hence unacceptable to
them . 

 

 That said windows and the internals esp registry have been changing
probably for exactly that reason.  Changing an OS from UCS-2 fixed char to
UTF-16 is not something you would do unless you had to.

 

Agree working with cross language runtimes is an issue that deserves more
consideration note however ASCII to UTF8 is a 1:1 but only for western
charactersets.

 

 

Ben


OK. This raises a practicum question. Is .Net legal in China?

I ask because .Net takes a middle position. While .Net strings can represent
the full Unicode code point space, the .Net char datatype is ucs2 (not
utf-16 - chars are strictly 16 bit entities). When dealing with code points
outside the basic plane, it is necessary to manipulate them as strings.
Java, I think, does the same thing.

>From the standpoint of runtime compatibility, BitC almost *has* to pick ucs2
characters. We can support ucs4 in parallel if we get the string issues
right.

So my question: is the .Net solution considered acceptable under Chinese
law? Whether it is or not, is it acceptable in actual practice?


 

_______________________________________________
bitc-dev mailing list
[email protected]
http://www.coyotos.org/mailman/listinfo/bitc-dev

Reply via email to