Hi Juan Pablo,
I am trying to answer your recurring question about displaying Braille output
of liblouis conversions.
I think you guess it well that display tables are the key to correct
representation of the Braille signs, although .dis files are not always needed
for that.
As I can see there isn't a separate chapter in the liblouis documentation
dedicated to display tables but the topic is covered in the sections dealing
with opcodes. display itself is an opcode. The trick is that display opcodes
may be gathered in a separate file and be applied at once prior to loading a
translation table.
The character-definition opcodes (especially the single-cell ones) implicitly
define a character representing the Braille dot combination associated to the
specified Braille symbol.
Once a translation table is completed, it will inherently have a set of output
characters representing the Braille cells that are the result of the
translation. These characters will appear in the output in place of the dot
combinations covered by the character-definition opcodes.
I don't have experience with too many translation tables, only with the en-us,
ueb, German, Hungarian and Esperanto tables. What these have in common is they
consistently cover the most commonly used input characters and have a default
set of output characters compatible with one of the three main Braille
representation standards: USA Braille, Euro Braille or Unicode Braille.
For instance, the en-us and UEB tables provide output according to the USA
Computer Braille standard, whereas the German and Hungarian translation tables
generate output by the Euro Braille standard.
Thus these tables produce a highly consistent output that contains only a range
of the 64 character codes allowed by the named Braille standard, in case of
6-dot Braille, or 256 different codes in case of 8-dot Braille. Of course, a
translation table cannot cover all possible unicode characters but in case of
these tables, apparently an effort has been made to cover the most frequently
occuring ones like the letters of the alphabet, digits, punctuation and the
most common other symbols.
A common characteristic of these translation tables is that their output will
not contain accented letters or both upper and lower-case characters appearing
in the same file but either upper or lower-case letters only. Characters in
place of punctuation and digit characters may or may not match their ascii
equivalents etc.
Once a text is processed using any of such translation tables, your screen
reader, Braille display or Braille embosser must be set to accept material
using the same character set.
The display tables may be used to alter the inherent, default output character
set of the translation table being used. Since for each Braille character
defined in a translation table, its Braille dot combination is defined by its
character-definition opcode, a display opcode targeting that dot combination
may be used to redefine the output character for that dot combination.
You can simply open a .dis file, these are plain text files and could be part
of translation-tables. Each line contains a display opcode that first specifies
the desired output character, then the Braille dot combination that is to be
represented by the character preceding the dot combination in the definition.
In case of a six-dot paradigm, such a file contains 64 display lines covering
all 64 dot combinations.
Thus a well defined set of display opcodes, practically put in a .dis file
creates an extra layer on top of the translation table. For instance, if I just
use a German table, like de-de-g1.ctb, the output will comply with the Euro
Braille standard, that is a specific set of 64 ascii characters representing
the 64 different Braille dot combinations.
But if my embosser is set to accept jobs in the USA Braille charset, then I
apply the us-table.dis display table prior to the German translation table and
then the table list will look like this:
us-table.dis,de-de-g1.ctb
Then the result will use the 64 ascii characters defined by the USA Braille
standard.
If I use the en-us-g2.ctb translation table, it will produce output already
according to the USA Braille character set, but if my Braille display is set to
Euro Braille and I want to read American Grade 2 material produced by liblouis
on the display without changing its settings, I can simply prefix the German
6-dot display table (that is the Euro Braille charset) to the en-us-g2.ctb or
en-ueb-g2.ctb translation table when running liblouis:
de-eurobrl6.dis,en-us-g2.ctb
A third common option is to produce output in Unicode Braille using the
unicode.dis file.
Thus the three useful display tables are these:
us-table.dis - American 6-dot Braille character table,
de-eurobrl6.dis - Euro Braille table originating from the German 6-dot Braille
table,
unicode.dis - Producing output using the Braille code block of the Unicode
standard. Most up-to-date screen readers and Braille displays can cope with
these character codes but many embossers cannot.
Now about the Spanish Braille translation tables in liblouis. First of all, I
am not an expert of Spanish Braille. I can read and approximately pronounce
Spanish texts written in grade 1 Braille but my poor knowledge does not extend
to grade 2 and my Spanish does not go beyond the usual patterns used by
tourists. So please treat anything I write here below in the light of my
Spanish expertise.
In the liblouis package (probably based on version 3.6), I found two main
tables I could practically test. One is es-g1.ctb, the other is Es-Es-g1.utb.
es-g1.ctb seems to have been created more recently (just a few years after the
another).
Es-Es-g1.utb produces output according to the USA-Braille table. Thus if your
screen reader/Braille display/embosser is set to Euro Braille, you may want to
use this translation table in combination with the Euro Braille display table:
de-eurobrl6.dis,Es-Es-g1.utb
Although Es-Es-g1.utb's file extension perhaps implies the use of Unicode, but
in my experience its output is purely 7-bit Ascii, and the other file,
es-g1.ctb is the one whose output contains certain Unicode characters.
I used liblouisutdml for the trials (I just saved the Espana main page of the
Spanish Wikipedia to an html file and fed liblouisutdml with it), and it seems
as if the character definitions in es-g1.ctb weren't covering most of the
commonly used characters, instead this table seems to deal only with special
Braille indications. Ordinary Spanish text remains as it is, including the
accented letters. These continue to be represented by their normal Unicode
characters and the same applies to punctuation characters. However Unicode
Braille capital signs are inserted before capital letters and Unicode Braille
number signs before numerals and digit characters get converted into letters to
suit the Braille rules.
I say "as if" because when I looked into es-g1.ctb, according to the include
statements in there, all the latin letters, including the Spanish accented ones
and punctuation should be handled by the table and I just don't know why this
is not happening.
However, this is a mixed output that does not contain Braille characters only
or by a correspondence table such as the Euro Braille or USA-Braille charsets.
Such output may be useful for individuals who are using their operating system
set to Spanish and their screen reader/Braille display set to interpret
ordinary Spanish text characters and besides that, their screen reader is
prepared to handle Unicode Braille characters. But this output may be
challenging for many embossers. Applying a display table to es-g1.ctb is also
largely ineffective.
Perhaps trying liblouis directly without liblouisutdml would yield a different
result but it would pose further challenges I tried to avoid. Namely, I should
refrain from using unicode in the command line but then in what code page
should a Spanish plain text be converted and how could I make sure the source
text contains the right character codes. It seemed easier to use a certainly
well-encoded HTML in UTF8 and let it through liblouisutdml.
But there may be a version problem involved. My windows build of liblouisutdml
can only work with the last December version of Liblouis. Perhaps that liblouis
version is no longer fully compatible with the latest Spanish tables. Thus the
second part of this message about my experiments with Spanish Braille may not
be relevant. Sorry for that.
But I hope the things said about the display tables above make some sense.
Best Regards, Norbert.
From: Juan Pablo Bello
Sent: Monday, November 19, 2018 11:33 PM
To: liblouis-liblouisxml@xxxxxxxxxxxxx
Subject: [liblouis-liblouisxml] Re: translating files to braille
wow, nice. I will be happy to translate the interface and aid with
music braille. However, to steer this back on topic, I have noticed
something that I just do not quite get, and its my main problem when
translating with liblouis.
When I create a brl or brf file and that does not interface directly
with my screen reader, I always get the wrong punctuation and symbols
for languages other than English. Does this have to do with display
tables? How are these created?
I am asking this because spanish does not appear to have a 6 dot brf
standard as English does and ... well, this presents major challenges
for forward translation. Curiously things do show up just fine when I
use NVDA though, why could this be?
2018-11-01 8:37 GMT-05:00, zvonimirek222@xxxxxxxxxx <zvonimirek222@xxxxxxxxxx>:
Hi Bhuc,
Can i translate this program to Croatian?
And, to test my translation,
Can you provide me a key?
-----Original Message-----
From: liblouis-liblouisxml-bounce@xxxxxxxxxxxxx
<liblouis-liblouisxml-bounce@xxxxxxxxxxxxx> On Behalf Of Dang Hoai Phuc
Sent: Thursday, November 1, 2018 1:34 PM
To: liblouis-liblouisxml@xxxxxxxxxxxxx; Juan Pablo Bello
<juanpisjaws@xxxxxxxxx>
Subject: [liblouis-liblouisxml] Re: translating files to braille
Hi,
The Sao Mai Braille is a kind of word editor and Braille translation
software and we are more than happy to provide you the free license keys for
non-commercial use. As we are a non-profit organization and we would also
need to generate certain income to maintain and develop more features for
it. New features are added such as dealing with music notation and Braille
music translation.
Regards,
Phuc
On 31/10/2018 4:46 AM, Juan Pablo Bello wrote:
well, as for the screen reader part...For a description of the software, to download it and links to project pages
I use unicode.dis to test translations interactively in windows from
command line, so no issues there. Or I use the es-g1.ctb table in NVDA
as output and it works as expected.
The Sao Mai Braille (SMB) program sounds like just the type of program
I would pay for and even contribute into translating (if that has not
been done already) so I will fiddle with it this weekend.
2018-10-30 16:37 GMT-05:00, Samuel Thibault
<samuel.thibault@xxxxxxxxxxxx>:
Juan Pablo Bello, le mar. 30 oct. 2018 16:23:25 -0500, a ecrit:
well, not really sure of what is going on, since when I translateBut how do you configure its braille translation exactly?
files using my screen reader this does get represented correctly.
Though yes, I notice a differente using lou_trace when testingWell, es-g1.ctb is made to use 6dots only, yes. If something else is
english tables. For example, a capital C shows 1. Uppercase C 147
where as with the es-g1.ctb table it just shows 1.
Uppercase C 14
desired then another table needs to be written.
Samuel
For a description of the software, to download it and links to
project pages go to http://liblouis.org
go to http://liblouis.org
For a description of the software, to download it and links to
project pages go to http://liblouis.org