Hi, Bert!
I solved some of the problems by using noback opcode. I'll provide tests that
show what's fixed and what's left to fix. I'll provide patches to fix a major
problem with the Norwegian tables that's going to be used together with
software like Handy Tech HTCom as well. HTCom uses liblouis to perform forward
and backward braille translation to Handy Tech note takers, and these devices
are not able to understand unicode braille patterns.
I'll come back.
Lars
On Fri, Feb 01, 2019 at 01:53:43PM +0100, Bert Frees wrote:
The fact that dots 56 and 345 have their own meanings when usedFor a description of the software, to download it and links to
separately at least seems to complicate things. I don't know if we can
get it to work somehow. As a start I think it would be good if someone
could verify that all these feature that we mentioned (the letter æ,
and marks to end computer braille, emphasis or capitalisation) are
covered by the tests, both in forward and backward translation. I've
just checked that the marks to end emphasis or capitalisation are
already covered by tests for forward translation.
Op vr 1 feb. 2019 om 13:33 schreef Lars Bjørndal <[1]lars@xxxxxxxxxxx>:
Hi, Bert!
Thanks for your suggestions. Let me follow up with some questions,
and let me start with the last part:
> If in addition you comment out the line "endcomp 345" you get a
correct
> roundtrip.
Yes, but I need the begcomp/endcomp respectively to mark phrases
like URLS or email addresses. Example: [2]nlb.no should be forward
translated into ⠣⠝⠇⠃⠄⠝⠕⠜. If removing endcomp, the end mark, dots
345, is missing. When I look into the documentation, however, I
cannot find endcomp/begcomp, and it seems that the compbrl thing has
changed significantly since we started to use it in the Norwegian
tables. I'd appreciate if you can shed some light on how to
accomplish liblouis to get URLS surrounded by start and end markers
nowadays. I see other tables are using these opcodes as well.
So to your first point:
> There is definitely a problem with the table.
> If you comment out the lines "endcapsword 56" and "endemph italic
56",
> the result of backtranslation becomes empty.
So you mean I cannot get both to work, ending block of capital
letters or italics, and at the same time get back translation of a
single letter æ to work?
Shouldn't that have been fixed in liblouis?
Thanks in advance
Lars
> Op di 22 jan. 2019 om 15:02 schreef Lars Bjørndal
> <[1][3]lars@xxxxxxxxxxx>:
>
> Hello list!
> I have a problem when I back translate a single character
> contraction.
> Steps to reproduce:
> 1. # lou_translate no-no-g1.ctb
> 2. Type `Dette er bokstaven æ' (without quotes) and press enter
> (means This is the letter æ)
> 2. Copy the translation and exit the program.
> 3. # lou_translate -b no-no-g1.ctb
> 4. paste the string and press enter
> I would expect that, since the letsign is used in front of the æ
> character, the back translation should not translate the letter
æ to
> the
> word være, which it actually does. Do I misunderstand someting,
or
> is there a proper solution to this
> problem?
> Thanks in advance - Lars
> For a description of the software, to download it and links to
> project pages go to [2][4]http://liblouis.org
>
> References
>
> 1. mailto:[5]lars@xxxxxxxxxxx
> 2. [6]http://liblouis.org/
For a description of the software, to download it and links to
project pages go to [7]http://liblouis.org
References
1. mailto:lars@xxxxxxxxxxx
2. http://nlb.no/
3. mailto:lars@xxxxxxxxxxx
4. http://liblouis.org/
5. mailto:lars@xxxxxxxxxxx
6. http://liblouis.org/
7. http://liblouis.org/
