[sword-devel] SIL Ezra fonts

Chris Little sword-devel@crosswire.org
Thu, 26 Sep 2002 14:09:25 -0700 (MST)

On Thu, 26 Sep 2002, Joel Mawhorter wrote:

> On Thu, 26 Sep 2002, Chris Little wrote:
> > SIL also released a new version of a nice transcoding utility to convert 
> > between different character encodings, which those of you venturing into 
> > module making might find useful.  See: http://www.sil.org/nrsi/teckit/
> ICU also has support for converting between Unicode and many many legacy 
> encodings. Since Sword already has some dependencies on ICU perhaps this 
> would be a better tool to use first. The only problem I see with using ICU 
> for conversion is if you want to convert from encoding a to encoding b 
> where neither a nor b are Unicode. In that case you have to do a -> 
> Unicode -> b, which may not give the same results as a -> b for some 
> encoding schemes. However, it looks like TCECkit has the same limitation.

This shouldn't be a problem for ICU since this is what compatability 
characters are for.  The only problem would occur if Unicode REALLY had no 
character for whatever both a & b could encode.

I wasn't suggesting using TECKit in sword, just that module makers might 
find it useful for recoding documents.  You can do this with ICU, if you 
want to fiddle with rules files and then recompile the data files.  TECKit 
can do the same without requiring a recompile.  Sword doesn't need to 
transcode since we only take data in Unicode. :)