Re: NNBSP-MVS Impact

Hi Greg Eck,
We (our team + Tugstuguldur) implemented Mongolian engine for Harfbuzz. 
The responsible person is Mr. Bahded.
I will contact you again, because just now I am on the trip to Mongolia.

Badral

On 02.07.2015 02:58, Greg Eck wrote:
>
> Does anyone have contact with a Harfbuzz engineer that can give us an 
> understanding on how Harfbuzz handles the character immediately 
> preceeding the MVS/NNBSP?
>
> Greg
>
> *From:* Greg Eck
> *Sent:* Tuesday, June 30, 2015 12:46 AM
> *To:* 'public-i18n-mongolian@w3.org'
> *Subject:* NNBSP-MVS Impact
>
> With reference to 
> http://r12a.github.io/scripts/mongolian/variants.html 
> <http://r12a.github.io/scripts/mongolian/variants.html>
>
> As we start the discussion of Mongolian variants, I think it is
>
> important to look at the means of implementing the NNBSP and the MVS
>
> first. There is considerable variation here. If we can understand how
>
> each font implements these two areas, it may bring understanding of
>
> why our fonts differ in shaping in these two contexts. I am presenting
>
> two documents for discussion dealing with the NNBSP and the MVS model.
>
> **If we could have a description of how Harfbuzz, Apple and other *
>
> *rendering engines interpret the character preceding/following the *
>
> *NNBSP/MVS that would be helpful.** These descriptions will be added
>
> into the NOTES section of each chart. The attached chart shows the
>
> model used in the Microsoft Universal Shaping Engine.
>
> Once we have agreed upon the two models, we will start with
>
> The U+1820-A (initial/medial )and then proceed to the
>
> other five similarly affected code-points U+1828 (medial/final), U+182C
>
> (medial/final), U+182D (medial/final), U+1835 (medial/final), U+1836
>
> (medial/final). A premise that I start with is that “Mongolian font
>
> implementation does not always follow grammatical rules” – case in
>
> point is the U+182C final which is not grammatical, but seems to be the
>
> implementation of choice for two glyphs in most of the fonts on display.
>
> Greg
>


-- 
Badral Sanlig, Software architect
www.bolorsoft.com | www.badral.net
Bolorsoft LLC, Selbe Khotkhon 40/4 D2, District 11, Ulaanbaatar