[liblouis-liblouisxml] Korean table: restructured ko files and fixed number-Ko char issue

  • From: Joseph Lee <joseph.lee22590@xxxxxxxxx>
  • To: liblouis-liblouisxml <liblouis-liblouisxml@xxxxxxxxxxxxx>
  • Date: Thu, 3 Jan 2013 13:12:51 -0800

Hi folks,
Update on Korean tables:
* Attached a zip file containing three files that are needed for
displaying Korean chars:
ko.ctb, which contains punctuations, numbers and Unicode sign opcodes
for Korean (Hangul) chars. This file also includes Korean dot/char
patterns which are not affected by grade 1 and grade 2 rules - that
is, entries which are common to grade 1 and grade 2. The "universal"
entries fixes an issue where some Korean conscenant dot patterns were
confused with litdigit (although in the table, I put after digit),
which was a concern raised by a number of Korean braille readers. For
example, the conscenant dots fpr "ni-eun" is dots 14, which are same
as number 3.
ko1.ctb, uncontracted Korean table. Includes har/dot pairs used in
Grade 1 rules. Mostly deals with "after digit" situations.
ko2.ctb, contracted Korean braille. Includes same chars as ko1.ctb
which has different dot patterns in grade 2. These are shorthands for
uncontracted chars. Also includes certain strings which must be
uncontracted and a number of multi-char ones which has shorthand dot
patterns.
* As for the structure of the files: ko.ctb is the base table from
which ko1.ctb and ko2.ctb derives. Thus, unlike other tables, ko2.ctb
does not include ko1.ctb. This would help me extend the table without
major modifications - all I need to do is categorize whether an entry
is universal or grade-specific, then add the entries in the
appropriate file (if it is universal, then ko.ctb will be modified; if
it is grade-specific, ko1 or ko2 will be modified).
BTW, for testers: these three files in the zip file has no errors when
checked against checktable (at least under 2.5.1).

On a related note: I found that if you need to use "always" opcode by
itself, you need to put it as the last entry for a char. For example,
I have a situation where a char might have a different dot pattern
when preceeded by a number, and other case where it does not. Then:
after digit always char @dotpattern1
always char @dotpattern2
Also, when you have a char that is represented in different dot
patterns, use "sign" first, then use "always".

In the meantime, I'll upgrade to LibLouis 2.5.2 and work on UEB tables.

For NVDA devs: Would you like me to wait a week before closing #2737?
//JL

Attachment: ko_liblouis.zip
Description: Zip archive

Other related posts: