Rick Jelliffe scripsit: > A control character may be a character or an embedded signal (i.e. a PI) > but it is certainly > not an element.=20 Of course not. But it may be *represented* by an element. > Furthermore, would we then have pre-control-expansion infosets and > post-control-expansion > infosets? (On top of the current pre-|post-[validation|namespace > processing|Xinclusion|XML Schema > augmentation] mess) No. This would be a universally available application convention, like xml:space and xml:lang, not affecting any infoset. > It would be better to reserve special characters which (like <) are > not allowed as literals, > for all the C0 and C1 controls. I don't understand this idea. You mean magic entity references? The trouble is that "<" is not actually magic, except that it needs no declaration: it has a definite replacement text. Something like "&BEL;" would have no legal replacement. > Or to allow numeric character > references, but that is less > tidy, because then people would be tempted to mark-up in code points > rather than in=20 > characters. Just so. -- John Cowan <jcowan@reutershealth.com> http://www.reutershealth.com I amar prestar aen, han mathon ne nen, http://www.ccil.org/~cowan han mathon ne chae, a han noston ne 'wilith. --Galadriel, _LOTR:FOTR_Received on Monday, 13 May 2002 13:12:50 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 29 October 2007 16:58:05 GMT