Re: Reviewed charmod fundamentals

On Mar 5, 2004, at 3:16 PM, Elliotte Rusty Harold wrote:

> Speaking as a Java programmer, I do not find UTF-8 to be less Java 
> friendly than UTF-16. Both UTF-8 and UTF-16 need to be passed through 
> a Reader on input and a Writer on output for any sort of robustness to 
> apply.  Which one I choose to use is almost never based on Java's 
> internal storage format for Strings.

I can't deny your experience, but speaking as a Java *and* C 
programmer, my experience is different.  I have the choice of 
generating text in my own language's idiomatic style, *or* of putting 
it through an output filter.  Output filters are not free and every 
cycle burned up in the XML-generation phase is being stolen from 
business logic.

One very small example: I recently wrote an XML-generation library for 
C (see
http://www.tbray.org/ongoing/When/200x/2004/02/20/GenxStatus) and in 
order for maximum usability by C programmers, it only knows UTF-8.  
Thus, it can't be used without extra filtering for any protocol that 
chooses to say "only UTF-16". -Tim