A few scripts are distinguishing sans-serif and serif letters (or two styles
for Arabic, Western "standard", and Eastern standard for Persan/Urdu).
This includes Armenian, Khmer, Lao... However the glyphs used for Latin letters
copied in them always use the Serif style instead of those from Noto Sans.
Merging fonts or copying Latin glyphs is proably not the best option. You can
just use the additional OTC table that gives the name of another linked font
for the Latin script.
This would reduce the size of fonts, avoiding duplicates and unsynchronized
updates between fonts (also this would avoid the issue where some Basic Latin
letters wil lbe taken from the local font, and most others froù an
unpredictable font).
Major OpenType engines can already process these font linking tables to build
"metafonts" with matching styles and metrics, preferably to all other metrics.
Avoiding duplicaes will improve performance, especially on small devie with not
a lot of memory and with multilingual documents (e.g. when loading Wiktionary
pages showing translations or pages containing language navigation links, even
if Wikiemdia solved the problem by creating a specific limited font designed to
show language "autonyms" for all supported languages, unfortunately that font
is still largely unhinted; may be you could as well provide a pan-script font
"Noto Autonyms" using a secltion of your hinted glyphs).
For now could you discriminate better between serif and sans-serif in fonts for
specific scripts that borrow a few Latin letters so they line up correctly?
Original issue reported on code.google.com by ver...@gmail.com on 16 Jul 2014 at 4:46
Original issue reported on code.google.com by
ver...@gmail.com
on 16 Jul 2014 at 4:46