[Scummvm-devel] Proposal for translation

[Alban Bedel]

On Sun, 17 Jun 2007 01:59:30 +0300
Eugene Sandulenko <sev at scummvm.org> wrote:

> On Fri, 15 Jun 2007 13:25:30 +0200
> "Jordi Vilalta" <jvprat at gmail.com> wrote:
> 
> > As I understand it, it seems the current GUI implementation isn't
> > capable of showing different character sets
> A thought. What about use of codepages, just like old Win95? I.e. each
> language will have font with special characters accompanying it. We
> have Latin-1 now, which is enough for most European languages. Then
> there could be Cyrillic font for Russian.
> 
> The problem with this approach is that it removes possibility to add
> Eastern encodings which usually use 2 bytes per character.

And it is a lot more complex that what you think. For ex. some characters
exist in several codepages but with different codepoint. This mean you
need either one font for each codepage or one decoder for each codepage
(to translate to unicode). Plus you can't (easily) display text with a
mix of various alphabets (think for ex. of a list of contributors) and
supporting alphabets that need multibyte chars become a pure nightmare.

On the other hand UTF-8 support all unicode char, is trivial to decode
and 100% compatible with the classic null terminated strings. Sure
rendering complex scripts like arabic or tibetan is not trivial, but
a simple render is still able to display (more or less correctly) a lot
of alphabets. And anyway the eventual complexity stay in the render
instead of creeping all over the code. This mean small targets can go
along with a simplistic render and those target that have the ressources
can have bidir support and all the mess needed for a really good rendering.

	Albeu