On Fri, Jul 17, 2009 at 6:13 AM, Tei<[email protected]> wrote: > I work for a service for multicultural education. Something I notice, > is that often the traductor uses a custom font. Theres a hebrew > custom font, a urdu custom font, etc... using these fonts, you can > write in these languages using unicode-unaware word processors. The > text is not really in unicode. If you copy&paste the code from one of > these word / pdf files to a unicode aware text editor, you get > gibberish, like "xfdLnfO{ yfxf 5 ;a} dfG5] leGgleGg :jefj / k|j[lQsf > x'G5g\ t/ klg zflGt /" (thats was a real text from a nepali > document).
Formerly everyone used language-specific text encodings like ISO 8859. Hebrew uses ISO 8859-8, for instance, and most Latin-based European languages use ISO 8859-1 or ISO 8859-15. Also, you had fonts (Wingdings is a prominent example understandable by normal people) that just substituted glyphs in the font file itself. This approach is obsolete and can be ignored for our purposes. MediaWiki only supports, and only should support, UTF-8 output. > IF there where a system to detect If a guy have a particular Hebrew or > Urdu font. That would open the door to.. just.. use > ..that..font..that...is ..already installed. The "Arial Unicode" > font, or the "DavidD" (hebrew) font, etc.. Browsers will usually do this already. If a character is not available in the font being used, they'll try to find it in some other installed font, and will only render a box or other error character if they can't. Firefox, at least, has behaved like this for a long time. I seem to recall IE didn't used to do this, but recent versions hopefully do. _______________________________________________ Wikitech-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/wikitech-l
