Braille translation software

Braille translation software

[Jason White]

Is there any braille translation software, for which source code is
obtainable that can be compiled under Unix/Linux, that is designed to
support a variety of languages as well as technical codes such as those
used to represent mathematical notation? I have examined the obvious
possibilities (e.g. NfbTrans and the Linux version of TurboBraille), but I
am not satisfied that, in basic design, they provide sufficiently general
translation mechanisms to enable rules for a broad variety of braille
codes to be readily defined.

Another interesting translator, though it has certain limitations, is that
developed by the RNIB: http://www.dpawson.co.uk/ which uses the
translation algorithm originally developed by Sullivan et al., in the
DOTSYS III translator (which later formed the basis of the commercial
Duxbury translator). The translation technique is interesting, owing to
its generality (almost every aspect of the braille code is defined in the
translation tables, including the context-dependencies and rule types).

I am sure that innovative braille translators have been developed in
Europe as well (including multilingual ones) and I know that research has
been carried out into the translation of technical (including
mathematical) material, but none of the source code appears to be
available at present.

I also support the Braifo project (http://braifo.sourceforge.net/) which
doesn't appear to be under development at the present time.

I would be pleased to collaboratewith anyone who is interested in the
development of better braille translators, and to offer whatever knowledge
and skill I can to the furtherance of this technology, taking advantage of
modern standards in markup languages, style sheets etc. One issue that
must also be borne in mind is the requirements of braille display drivers
such as BrlTTY and the limited memory and other resources of portable,
Linux-based devices; an ideal translation system would consist of a
modular library for both translation and formatting that could be run on
portable devices while taking advantage, in conjunction with scripting
languages, style sheets etc., of modern document markup standards.

Any relevant information would be much appreciated.

Jason White.

Braille translation software

[Ben van Poppel]

This is very interesting, and something I've been thinking about a
fair bit over recent months, being both a reader of BUOC
(user-oriented extension to standard gr2 braille) and contracted
German Braille (Kurzschrift).

It would be good to see a translator produced that had more the look
and feel of the standard GNU utilities. I had an idea a while ago
about a table system that used regular expressions for matching,
thereby providing a very powerful mechanism for searching for strings
in a wide variety of contexts. I even went as far as the beginnings of
a sketch for a (possibly Python) prototype, but it is far from
complete. 

Regards,
Ben

Re: Braille translation software

[Jason White]

There is in fact an article published in the Perl Journal describing a
Perl script which took the rule file from NfbTrans, converted it into one
large regular expression and executed it. The accuracy was quite
reasonable though by no means perfect. I am not sure what the memory
requirements were or what the limitations would be on the size of the
resulting regular expression. The full requirements of braille translation
would probably demand a somewhat more complex approach: a pattern language
allowing rules to be conditional upon the application of prior rules,
provision for proper handling of Unicode, case distinctions, etc. A
reasonably competent computer science expert with knowledge of
finite-state techniques would be able to design and implement such a
system without much difficulty. Needless to say, I don't qualify under the
above-mentioned criterion, but I would be entirely willing to work with
someone who does.

Re: Braille translation software

[Luke Davis]

May I request that people CC the Linux-Braille list on this topic?  The
address is:

LINUX-BRAILLE Distribution List <linux-braille@braille.uwo.ca>

Thanks

Luke


On Thu, 3 May 2001, Ben van Poppel wrote:

> This is very interesting, and something I've been thinking about a
> fair bit over recent months, being both a reader of BUOC
> (user-oriented extension to standard gr2 braille) and contracted
> German Braille (Kurzschrift).
> 
> It would be good to see a translator produced that had more the look
> and feel of the standard GNU utilities. I had an idea a while ago
> about a table system that used regular expressions for matching,
> thereby providing a very powerful mechanism for searching for strings
> in a wide variety of contexts. I even went as far as the beginnings of
> a sketch for a (possibly Python) prototype, but it is far from
> complete. 
> 
> Regards,
> Ben
> 
> 
> 
> _______________________________________________
> Blinux-list mailing list
> Blinux-list@redhat.com
> https://listman.redhat.com/mailman/listinfo/blinux-list
> 

Braille translation software

[RAYNER Peter]

I started looking into this a few years back with lex/flex/bison but,
like many personal use projects, 95% right was good enough for me and
the last 5% looked jully hard.  I'd now have to relearn all the
grammar rules.  Is a stepping stone to this to wonder whether formal
grammars can be produced to represent the various braille codes?  If
we can do this there's a vast array of machinery I suspect we can turn
on the problem.  
cheers
Peter

Re: Braille translation software

[Jason White]

As I understand it, there are different kinds of formal grammar, each
with its own capabilities and limitations.

The language that was proposed for the Braifo project
(http://braifo.sourceforge.net/) is quite good. So far as formatting
is concerned, it would be useful to have a system that was independent
of any particular style or markup language (perhaps a function library
that could be called from a scripting language or an application
program to control the format, leaving the processing of the input
document, whether as a text file, an XML document instance, a LaTeX
file, etc.), to a higher layer--namely a script or application.

Finally, it would be useful to write conversion scripts that could
take the table formats used by other translators (E.G. NfbTrans,
TurboBraille, the RNIB ITS system, etc.), and convert it to the
language used by the proposed translator.

Also, I was recently alerted to the existence of an XSL style sheet
for producing the French mathematical braille code from MathML input.
Unfortunately it requires another translator to process the textual
portions of the document, but the XSL approach is nonetheless
interesting: see http://handy.univ-lyon1.fr/projets/bramanet/