PUA == "Private Use Area", so people can show whatever glyphs they want for whatever PUA code point they want. It's more like per-font or something than per-locale. Different documents could use different fonts to show different things.
We map those to the Unicode PUA, there's no better Unicode code point. FWIW: We have a mechanism where we allow "EUDC" characters to be mapped. The net result is that people can cause a specific font, of their own creation, to be used as the fallback for the system for those unknown PUA characters. For a web site, that'd mean that if they wanted to use the PUA, they'd either have to use a common convention, or provide a font. In either case I'd strongly recommend that the web site developer used Unicode as, particularly in these edge cases, the differences between implementation make it really hard to be cross-platform. -Shawn -----Original Message----- From: Anne van Kesteren [mailto:[email protected]] Sent: Wednesday, March 28, 2012 9:19 AM To: Shawn Steele Cc: [email protected] Subject: big5 and big5-hkscs Hi Shawn, I was hoping you could clear something up for me regarding big5. As far as I can tell Internet Explorer treats big5 and big5-hkscs the same. When generating all the possible multi-byte sequences (0x81 to 0xFE as lead, 0x40 to 0x7E and 0xA1 to 0xFE as trail) I get 19782 code points of which 6217 are in PUA in Internet Explorer. Is it still the case (as suggested by http://www.microsoft.com/hk/hkscs/ and elsewhere) that these map to different glyphs depending on the user's locale? Other questions you could really help me with: 1. If they do indeed map differently, is there a way to get more information as to how they map differently? 2. Is there information available what the best Unicode code points for these PUA code points are? I did not email this to the charset list as it seemed off-topic. I did however cc www-archive as this information might be relevant to other people. Hope that's okay. Kind regards, -- Anne van Kesteren http://annevankesteren.nl/
