- From: Julian Reschke <julian.reschke@gmx.de>
- Date: Thu, 3 Oct 2002 16:58:24 +0200
- To: <www-tag@w3.org>
Hi, I think we need to consider the following questions... If a "standard" mapping from QName to URI is defined, .... 1) does it need to be unambiguous? 2) does it need to be reversible? 3) does it need to work with all legal namespace names, or is a subset enough? 4) should it allow URI references as mappings? 5) how will it treat namespace names in XML 1.1 (assuming that the extension to IRI references is there to stay). Some toughts: 1) Yes. 2) If it's not, it's of limited value in many cases, because you then can't round-trip the QName. Note that the mapping proposed in [2] does not have this property. 3) As namespace names can use *any* URI scheme, limiting the mapping only to some URI schemes seems to be problematic. However, if all URI schemes need to be handled, generating a URI in the same scheme as the namespace name simply won't work (because the URI scheme may not allow *any* kind of extension/concatenation). 4) If the namespace name already has a fragment identifier, adding another one isn't going to work, right? 5) This would require an unambiguous mapping from IRI references to URI references. Is that defined? Julian [1] <http://www.w3.org/2001/tag/ilist#rdfmsQnameUriMapping-6> [2] <http://www.w3.org/2002/09/24-tag-summary#rdfmsQnameUriMapping-6> -- <green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760
Received on Thursday, 3 October 2002 10:59:01 UTC