Re: Mongolian NNBSP [I18N-ACTION-458]

On Sat, 1 Aug 2015 14:44:52 +0900
<jrmt@almas.co.jp> wrote:

(I've changed the quoted email to give the cases letters.)

> 1.     The NNBSP word break property should be changed to MidLetter
> to fit Mongolian Suffix usage.
> 
> (a)  MLetter x NNBSP x MSuffix                   : MLetter ->
> Mongolian Letter, MSuffix -> Mongolian Suffix
> 
> (b)  ALetter x NNBSP x MSuffix                    : ALetter ->
> Alphabet Letter
> 
> (c)  Numeric x NNBSP x MSuffix                  : Numeric (digit and
> Mongolian digit)
> 
> (d) MLetter(Extend|Format) x  NNBSP x MSuffix : Extend | Format ->
> FVS1-3, ZWNJ, ZWJ etc)
> 
> (e) Some_Punctation x NNBSP x MSuffix   : Some_Punctation like )>]” ‘
> etc.  (If this is not possible to define separately, we can replace
> it with following isolate Mongolian Suffix)
> 
> (f)  NNBSP x MSuffix                                   : We need to
> support the isolate Mongolian Suffix exist

I'm assuming that MSuffix always starts with a letter.  There seems to
be no need to make Mongolian letters a special type of letter - ALetter
suffices.

MidLetter doesn't work with (c), nor does MidNumLet.  The UTC is
offering ExtendNumLet, which will cover all but (e).

Nothing covers (e).  For example, English "word(s)" is split so that
the words are "word" and "s".

Although it is covered, is (f) suitable for the start of a line?
I'm not sure that the NNBSP can be rendered properly in such cases.
Would "ZWJ MSuffix" suffice where "MSuffix" was insufficient?

Should NNBSP ever be at the end of a line?  If it is given a general
category of Pc rather than Zs, it is likely to be displayed.  Perhaps
this does not matter.

Richard.

Received on Saturday, 1 August 2015 08:39:13 UTC