- From: David Carlisle <david@dcarlisle.demon.co.uk>
- Date: Sat, 10 Jun 2000 11:04:32 +0100 (BST)
- To: jjc@jclark.com
- CC: XML-uri@w3.org
James, thanks:-) a version of the proposal that I can understand! > 1. A1 is character for character identical to A2, and > > 2. either > (a) A1 and A2 are absolute, or > (b) both > (i) A1 and A2 are relative, and > (ii) C1 and C2 are character for character identical > > Is this what you have in mind? > This version would be fine by me. It avoids the main (I would argue fatal) problem with the "make absolute" approach of effective element names changing in mysterious ways, and as far as I can see it would not break any of my existing documents using relative namespace names. Actually effective element names would still change as documents were moved, moving <x xmlns="a"/> from C1 to C2 would change the effective element name from (C1,a,x) to (C2,a,x) but the bad effects of this are very much reduced (to becoming insignificant) because having effective names being a triple rather than a pair means that the name doesn't change to being equivalent to the name of an element which had a different string specified in the attribute value. David
Received on Saturday, 10 June 2000 05:59:46 UTC