RE: UTF-8 <> UCS-2/UTF-16 conversion for library use

From: Carl W. Brown (cbrown@xnetinc.com)
Date: Mon Sep 24 2001 - 16:32:22 EDT


Mike,

>
> "If you think you have the answer to all the problems, then you
> don't know all the problems."
>
> I tried to make a point, and apparently made it poorly. I will try
> again. It seems that some people are arguing that UTF-16 is the ideal
> solution for all computing, and that UTF-8 and UTF-32 exist only
> for network
> transport. Without getting bogged down in details, I care to disagree.
>
> Analyze problem. Pick solution. In that order.

I think that most people will agree. However I think that I do not
recommend UTF-32 networking because of the big/little endian problems. It
also is slower to transmit.

The selection of UFT formats used within the application depends on many
different factors. I have found that some applications benefit from a mix
of different formats. This is why I implemented xIUA to support a parallel
set of routines for each UTF format.

Carl



This archive was generated by hypermail 2.1.2 : Mon Sep 24 2001 - 15:20:57 EDT