[Gdal-dev] RFC DRAFT: Unicode support in GDAL

[Mateusz Loskot]

Armin Burger wrote:
> When I remeber well, you have to specify globally in the Windows 
> registry if the .dbf of shapefiles are encoded in UTF. But if you do 
> so and then load a shapefile not in UTF then the special characters 
> are displayed wrong.
> 
> So ArcGIS has not really a flexible support for UTF, at least not for
>  shapes.

Yes, you're right, although this setting applies to ArcGIS behavior.
AFAIU, it's something like turn on/off Unicode support :-)

If we want to work out more flexible solution we need to know what's the
data encoding itself, because in fact we will deal with 3 encodings:
local on user's computer, UTF-8 used by GDAL and dataset encoding.

IOW, we need to know how to read data and to convert it to UTF-8 in
order to push it to user through another conversion from UTF-8 to user's
locale. This is my understanding of the problem.

So, Shapefile's localization driver ID or .cpg file can be used here,
regarding OGR and Shapefile driver only, as an example.


Cheers
-- 
Mateusz Loskot
http://mateusz.loskot.net