As usual, I found the answer at Maha Anna's discussion page:
Java uses a system called UTF for I/O to support international character sets
True. Java uses a conversion method called UTF-8 which is a subset of UTF. Subset in the
sense, in true UTF a char can be encoded from 1 byte to ANY no of bytes. This means we
can cover ALL CHARS IN ALL LANGUAGES IN THE WORLD. So UTF is true transformation.
This means a small char can have lesser no of bytes , at the same time a BIG-LOOK&FEEL
( ) asian char may be encoded with many no. of bytes. Since in Java all chars can have
max 16 bits,(Unicode char) , All IO operations which need char transformation of bytes (All
readers/writers) uses a pre-defined transformation format. (i.e) a char can be encoded to
1 or 2 or 3 bytes ONLY . max 3 bytes. There are some rules which chars are encoded with
how many no of bytes. It is in Java
Doc. I also found a error in Bill brogden's book recently
here which illustrates this concept.
the link
www.javaranch.com/maha/Discussions/java_io_Package/true-false_-_JavaRanch_Big_Moose_Saloon.htm [This message has been edited by Tom Tang (edited February 12, 2001).]