Re: Mongolian NNBSP [I18N-ACTION-458]

On 31 July 2015 at 15:58, Badral S. <badral@bolorsoft.com> wrote:
> On 31.07.2015 13:46, Richard Wordingham wrote:
>>
>> So far as I am aware, Mongolian words do not start with NNBSP.
>
> Correct.

However, users may wish to illustrate suffixes in isolation (e.g. for
educational purposes).  This could be done with the help of ZWJ, but
it would be natural for a user to copy a Mongolian suffix (i.e.
NNBSP+suffix) from Mongolian text and paste it in isolation (e.g. at
start of line or after Latin text) and expect the suffix letter shape
to be preserved.

>> The relevant sequence is Aletter NNBSP Aletter, for which see
>> Examples 6 to 9 in the referenced link.  The first example is the
>> *three* characters "c.d", in which by the rules there is no word break.
>
> Not entirely correct. The possible sequences are:
> Aletter NNBSP Aletter, FVS1 NNBSP Aletter, FVS2 NNBSP Aletter, FVS3 NNBSP
> Aletter, Numeric NNBSP Aletter, ...

FVS1-3 have a word break property of "Extend", which according to Word
Boundary Rule 3b <http://unicode.org/reports/tr29/#WB3b> means that
they would inherit the ALetter property preceding Mongolian letter.
Therefore there is no issue with FVS preceding NNBSP.

What is the context for "Numeric NNBSP Aletter" ? ExtendNumLet would
inhibit a word break after a numeric, but I think that MidLetter would
not.

Andrew

Received on Friday, 31 July 2015 15:45:21 UTC