RE: Char node-type

[Kay Michael <Michael dot Kay at icl dot com>]

> Converting the characters in your source XML into 'images' that can be
> rendered in an HTML browser is a challenge in its own right.  We have
> identified three types of challenge:
> 
> - well-defined Unicode characters which a 'typical' [add your own
> definition here] client browser cannot be relied upon to render;

Yes, we developed a multilingual site for a client (www.justask.org.uk) and
found it quite possible to do English, Welsh, and Cantonese using Unicode
fonts, but after a lot of experimenting we gave up on Urdu/Bangali/Punjabi,
using GIFs instead. Whether the problem was with the browsers, or the font
suppliers, or (as our translators suggested) with the Unicode repertoires
for these scripts, we never really established. 

But we don't attempt to do a one-to-one character-to-GIF mapping, instead we
use an offline build process to assemble composite GIFs that represent whole
chunks of text. Of course that could in principle be done using XSLT as
well. Both approaches are slow, but which is better probably depends on
caching behavior, on how volatile the data is (ours is very stable) and on
the ability of your server to handle the high number of requests.

Mike Kay


 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list