Would you
test these, please?
I've created a filter to replace all instances of a
named entity with its numerical counterpart based on a given DOCTYPE.
The current information flow is:
Bookeen dict -->Using Penelope--> Stardict-->using script --> xdxf
xdxf -->using script-->Pocketbook dic
xdxf -->using Linguae with the settings unsorted and sametypesetting h -->Pocketbook dic
(Linguae makes a mess of this dictionary if it tries to ASCII sort and searching for 'a' would then return 'ABN'.)