On Apr 8, 5:05=A0am, Phillip Gawlowski <cmdjackr... / googlemail.com>
wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Austin Ziegler wrote:
>
> | (Yes, Virginia. Most people don't need full-on Unicode munging in
> | their code. It's necessary when you do need it, but most people don't
> | need it.)
>
> http://www.joelonsoftware.com/articles/Unicode.html
>
> I need it. Most of Europe needs it. Not to mention Arabia, Japan, and
> everybody else not speaking English.
>

Bullshit. I've done localization for Chinese sites in Ruby without
ever
having to think about unicode issues. SOME things require you to take
care. Such as not making stupid assumptions about the length of
strings
in bytes vs. characters. But most of the time people can do a lot of
string mangling with UTF-8 without ever needing to think about it.

ALL my work involve international character sets in one way or the
other,
because I refuse to not deal with i18n. NONE of my work involves "full-
on
unicode munging". It involves a tiny subset of support at most.

That applies to most people, and most of what they'll ever need is
trivial to implement.

Reading/writing UTF-8 is easy. Splitting UTF8 into characters, or
converting it to 32 bit codepoints is easy (surprise, surprise,
#pack and #unpack supports it).

Once you have it in 32 bit, the length calculation problems go away,
and slicing and  splicing strings without thinking about it becomes
easy.

That's almost everything _most_ people ever need in order to work
with Unicode. If there was such a burning need, most people would
use ICU bindings for Ruby (do they exist? I've never even bothered to
look, but if they don't they'd be trivial to write if I ever need
them)

Vidar