On Sun, 16 Feb 2003, Doug Ewell wrote: > The Unicode home page includes the following line, right where it should > be, in the <head> section: > > <meta http-equiv="Content-Type" content="text/html; charset=utf-8"> > > Any User Agent that takes a page properly marked as UTF-8, as above, and > still tries to autodetect a local code page, is badly misguided. How > would it handle a real UTF-8-encoded euro sign (0xE2 0x82 0xAC)?
AFAICR, there is supposed to be no single non-ASCII character before that <meta> tag. I really don't like to search the specs again, but I'm sure I saw it somewhere. The HTML renderer sees those characters and thinks the document has already started (since the <html>, <head> and <body> tags are are not mandatory in HTML 4 Transitional). So it goes into autodetection mode. The same situation happens with MS FrontPage 2000 (but I've already explained that). roozbeh
