Re: .webfont Proposal 2

This seems to me a significant step forward, Tal and Erik. It is helpful 
to see what this kind of .webfont might look like.

Karsten and I have been talking about this in terms of the composite 
font standard process we're involved with, and I think you might be 
interested in implementing something like the XML syntac Ken Lunde 
(Adobe) has proposed for identifying character support in component fonts:

	<Encoding Target="0028-0029, 002C, 002E,
	0041-005A, 0061-007A, 2018-2019, 201C-201D">

This seems to me much more reliable than the Unicode Range bit data that 
EOT and EOT Lite include, or the corresponding data in the OS/2 font 
table. The criteria for claiming Unicode Range support in a font is 
deliberately vague, so the information is not a good indicator of what 
characters are actually supported by a font. [OS/2 codepage support bits 
are generally more reliable, but fail to identify supported characters 
outside of standard 8-bit charsets.] The only way to reliably identify 
character (and hence script and language) support in an OpenType font is 
to parse the cmap table.

Having simply expressed character range information in the info.xml file 
seems to me a significant benefit of a wrapper format.

John Hudson

Received on Wednesday, 15 July 2009 18:49:42 UTC