Chris Smith writes:
Andreas Gohr wrote:Preben Randhol writes:
Is there any technical reasons why one don't simply use the titlename? I mean one can convert it with rawurlencoding and replace spaces with underscores, but I don't understand why one need to force ASCII. Many pages will be written in UTF-8 so why not keep it?
there was a longer discussion about it some time ago on the list. as far as i can remember IDs may only contain ascii characters, underscores and numbers (specified in the XHTML specs). They are also not allowed to start with a number.
Andi, that is the html restriction. xhtml follows xml spec for ID, refer http://www.w3.org/TR/xhtml1/#C_8 the third para in particular. That effectively means just about the whole of unicode is acceptable. You'll find the formal definition in the XML specs. From a validity point of view, there doesn't seem to be any reason not to use the 'deaccent' setting to determine the fragment id.
Andi -- DokuWiki mailing list - more info at http://wiki.splitbrain.org/wiki:mailinglist