- From: Michel SUIGNARD <Michel@suignard.com>
- Date: Tue, 2 Jun 2009 22:10:13 -0700
- To: "Michael Day" <mikeday@yeslogic.com>
- Cc: "John Daggett" <jdaggett@mozilla.com>, <www-style@w3.org>
>From: Michael Day > >Using script values sounds like a great idea! That's a lot easier for >the user than specifying a dozen Unicode blocks, and it also handles the >common and inherited characters, which unicode-range currently can't do. > >(It's also nice from an efficiency point of view, as we already have >perform script processing in order to correctly apply OpenType features, >so it is no extra work at all). > >So, how about allowing unicode-range to accept Unicode script names? And >should these be strings or keywords? :) Hi Michael, For the names if you looked at the Unicode UAX#24 (latest is http://www.unicode.org/reports/tr24/tr24-11.html you'll see that you can either use the long names (such as Cyrillic) or short 4 letters names (such as Cyrl), ref the sc entries in http://www.unicode.org/Public/UNIDATA/PropertyValueAliases.txt I guess the 4 letters are more like keywords, while the long names are more like strings. Because the script value is unique for each Unicode character, a script set is equivalent to a set of Unicode ranges, so they could be used interchangeably with Unicode ranges in the syntax. The problem though occurs with the Inherited and Common sets, although they are required for many writing systems, you would not expect a font to contain all of them. In fact you would only expect a subset required to support the other scripts specified by a given font. In other words, you would explicit declare a set of scripts in the @font-face syntax, and make sure your font implicitly contains the appropriate set of Common and Inherited characters to adequately represent the writing systems covered by a given script (or set of scripts). This requires some detailed analysis. Michel
Received on Wednesday, 3 June 2009 05:10:49 UTC