> I don't have the answer, but unfortunately it seems to me that at some point > visual clues (be they flags or bitmaps of the "text" with appropriate > alt > text rendered in Roman scripting) would have to be the most accessible > solution (pragmatism vs. standards zealously). Thoughts? I could see most, though not all, of the characters on that page. The reason I could see them is I have a font installed (Code2000) which sacrifices quality to enable it to pack in a large number of glyphs. While this won't be a high-quality representation it is a useful fallback behaviour. Apple have a Last Resort font <http://developer.apple.com/fonts/LastResortFont/>. It offers a fallback behaviour for characters no glyph can be found for by displaying a glyph which represents the Unicode block it came from - which at least gives you a clue in finding an appropriate font to display it properly. With a font like Code2000 (and Code2001 which goes beyond the BMP) multi-lingual text should at least be legible. Okay we need to get those fonts onto users machines first, but I think that's going to happen sooner than the machines reading text in images. Another possibility is font-embedding, embedding would allow you to provide a font that contains glyphs for all the characters you used. However I'd prefer to be able to embed a font to be used only the the users personal choices couldn't cope with the text in question, that I think would be the best-of-all-possible-worlds solution.Received on Wednesday, 1 October 2003 12:54:10 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 19 July 2011 18:14:11 GMT