[ros-kernel] l18n in react os
KJK::Hyperion
noog at libero.it
Thu Oct 9 02:05:42 CEST 2003
At 02.16 09/10/2003, you wrote:
>áéåí øáéòé, 8 áàå÷èåáø 2003, 21:39, ðëúá òì éãé Rick Parrish:
> > Diego Iastrubni wrote:
> > >Hi all,
> > >
> > >I was wondering what will be the approach in this project for i18n.
> > >Is it possible to use gettext()?
> >
> > Actually, that's not strictly true - although a lot of developers choose
> > to do it that way. You can have multiple localized resources or separate
> > resource DLLs for each locale. The approach is to be as compatible as
> > possible - UTF-16 and MBCS (UTF-8).
>i was thinking something similar.
>how about dynamic layout?
a-ha! the same instant I read:
>>>The approach in MS world is to recompile with new strings.
I expected to read something about "dynamic layout". Well, Windows dialogs
don't support dynamic layout, period. If you look for dynamic layout in
dialogs, you've misunderstood them: dialogs are simple and dumb, and
supposed to. But since you certainly *don't* want your main window to be a
dialog, the problem doesn't exist
>try making a dialog for english and then translate it into german. not all
>words will be visible since german translations are most of the much
>bigger then english ones.
you can't add dynamic layout to dialogs without causing an unstoppable
chain reaction of third-party application breakage. Want to write your own
layout engine anyway? you're welcome, just throw it on top of that
mile-high pile of layout engines over there, when you're done
>I will not even start talking about rtl and bidi yet.
Windows supports RTL layouts perfectly: just set the appropriate style bit
for a window, and everything in the window's DC gets automatically
mirrored. In most cases you don't even need to touch your code. BIDI?
recent versions of Windows have added support for complex scripts - you
could even write in alternate ltr-rtl mode, or in vertical. We don't have
to invent anything: like usual, those kind folks at Redmond have already
done all the boring research for us
More information about the Ros-kernel
mailing list