From: Peter Constable (petercon@microsoft.com)
Date: Wed Nov 21 2007 - 14:24:50 CST
> From: unicode-bounce@unicode.org [mailto:unicode-bounce@unicode.org] On
> Behalf Of John Hudson
> The implementation issue for embedded Mongolian is confused by the fact
> that the script is
> currently implemented in OpenType fonts as if it were a LTR script, i.e.
> the outlines are
> designed left-to-right in the glyph space rather than top to bottom,
> and then the text
> blocks are rotated for vertical display. This is itself a questionable
> practice, although
> it obviously simplifies font development and line layout for Mongolian.
I'm not sure why it would be considered questionable practice: There's no a priori requirement for how outlines need to be oriented in the design space, the history is not a determining factor, and it makes sense for the way that it is laid out in typical scenarios: drawing a line of Mongolian text in vertical layout or horizontally embedded in Chinese or English are straightforward (modulo the contextual processing that would be needed regardless). If the glyph outlines were designed in the 180°-rotated orientation from this, then all of these scenarios would require each Mongolian run to be drawn rotated 180°. If the glyphs were drawn rotated 90° clockwise from this orientation, then a CJK-like transform would be needed for vertical layout, rotating each glyph 90° counter-clockwise.
For people whose minds are pre-conditioned by knowledge of the historical derivation and familiarity with related RTL scripts such as Arabic, it may seem strange at first (as it did for me). From an implementation perspective, though, it makes good sense.
Peter
This archive was generated by hypermail 2.1.5 : Wed Nov 21 2007 - 14:26:27 CST