Re: The face of Unicode support in the future

On Wed, 19 Jan 2005, [ISO-8859-1] Florian Growrote:

> But I don't know what will happen when you replace (a part) of a String 
> with a String with another encoding. Or what happens when you 
> concatenate Strings with different encodings and so on...

I've been trying to follow this discussion, and since I know little
about character encodings, I am now hopelessly confused.

So I'd like to ask this, in the hope that the naivety might actually
stimulate useful thought:

Why can't we do for string encodings what we do for numeric
encodings?  Why not epresent things in the smallest type possible,
promoting to bigger, more sophisticated types as necessary,
automatically, and provide to_X methods to "explicit cast"?  I think
the difficulty is that bytes can encode all binary patterns, but can
we make use of the Byte Order Mark to force promotion?  When joining
strings that should only occur at the start of the whole string.

I'm fairly used to the idea with numbers, be they integer, float, or
whatever, though I sometimes err, but I caould then carry the ideas
over to strings of various flavours.

I'm thinking this because I'd like my ruby code to the right thing
in general without having to manage all the complexity myself, just
like I don't have to for memory management.

Now I'll stop "meddling in the affairs of wizards"

         Thank you,
         Hugh