Re: [gui-dev] Sorted list of languages.

"Ryan Keppel" <[email protected]> wrote:
>I assume the non-ASCII languages are sorted by UNICODE value? That should 
>be
> fine for the vast majority of the market.

See the translate page (http://www.limewire.org/translate.shtml), and you'll 
see that languages are listed by category for now, according to the charset 
they need (or support on various platforms), and then sorted alphabetically 
in each group. This is more or less by Unicode value, but for practical 
reason, it is easier to find this way, as this more or less corresponds to 
the default UCA order.

The language selector that will be implemented however will allow selecting 
languages by continent/region/country because many people don't know the 
name of their own script.

> If you think sixteen bit values are difficult,

They are not difficult. They are the standard for the Java *native* char 
datatype, which is NOT really a character but simply an unsigned 16bit 
integer. In Java, the correct model for characters is the "Char" class, 
which is already ready for handling supplementary planes.

> wait until you see the supplemental values (above 0xFFFF). Note to
> creators of Java++

Look into Java 1.4.1 and now 1.5. They are already supported.

> Make the char an unsigned thirty-two bit value and leave
> room for most of the galaxy's natural languages.

This was not needed in Java: the correct level to handle international text 
is the full TEXT level, i.e. "String" instances, not "Char" instances. They 
are immutable so their underlying encoding as arrays of characters or code 
units does not matter. The standard Java "String" class is already ready to 
handle characters in supplementary planes.


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