Hello Meikel!

Replying to a message of Meikel Brandmeyer to Neil Heller:

 NH>> My concept of localization requires the use of a string table.  The 
 NH>> impression I got was that sharable resources are slower to load and 
 NH>> execute than their counterpart (strings written into source code).
 MB> It is my impression, that string lookups more flexible than loadable
 MB> shared resources, ie. one can change the localisation at runtime,
 MB> whereas those shared resources are "hardwired". Or am I getting
 MB> something wrong.

There is more than one way to create translated software.  Regardless, you
cannot hardcode the text in your source code.

One way is to create a DLL (or cat file on unix) that you load at runtime
to get your string resource from.  Another is to bind the string resource
directly to your code, and you create different versions of your software
for each language (only on Windows and OS/2).

The latter option should be almost as fast as hardcoded text - there's no
extra file(s) to load.  But, to be honest, all "output" is slow -
even painting text to a GUI screen.  The overhead in getting the right
translated text vs actually using it won't be noticed on any computer that
is still around.

Darin

---