[liblouis-liblouisxml] Re: AW: Re: 2 questions: 1. charsets for output, and 2. Python bindings for hyphenation
- From: Christian Egli <christian.egli@xxxxxxxx>
- To: <liblouis-liblouisxml@xxxxxxxxxxxxx>
- Date: Fri, 12 Feb 2010 09:03:30 +0100
"Leo" <fhaxbox66@xxxxxxxxxxxxxx> writes:
> we wonder if lou_hyphenate will produce 100 per cent correct results
> in all languages. It seems that German has wierd hyphenation rules for
> g2 Braille. I am not too familiar with them, and all this does not
> affect the usefulness of lou_hyphenate. My only question is whether it
> would make sense in the long run to look for something different,
> maybe outside liblouis. But this is off topic.
The problem with hyphenation and contracted Braille is that the quality
depends on the back-translation, as liblouisxml (when doing line
spitting) first translates a contracted back to uncontracted, then
hyphenates and then contracts again. So if the back-translation is bad your
hyphenation will be bad. Another and probably better approach IMHO would
be to create dedicated hyphenation tables for contracted Braille. This
would give much better results as it doesn't depend on back-translation.
Christian Waldvogel has some pointers as to which tools to use to
create a hyphenation tables from a corpus if you're interested.
Aside from that it might be beneficial to use a dedicated library for
hyphenation if there is one. hunspell might fit the bill but I haven't
looked into it too closely.
Thanks
Christian
--
Christian Egli
Swiss Library for the Blind, Visually Impaired and Print Disabled
Grubenstrasse 12, CH-8045 Zürich, Switzerland
For a description of the software and to download it go to
http://www.jjb-software.com
Other related posts:
