- From: Eric van der Vlist <vdv@dyomedea.com>
- Date: Wed, 24 May 2000 20:20:29 +0200
- To: Tim Berners-Lee <timbl@w3.org>
- CC: xml-uri@w3.org, "Simon St.Laurent" <simonstl@simonstl.com>
Tim, Tim Berners-Lee wrote: > > XSLT uses XPath which is included, I understand in the "lower layer" in your > scenario. I am not sure I fully understand your example :( The problem I see with transformations, is that you can define 3 different XBases : 1) the base for the nodes of the document tree. 2) the base for the nodes of the XSLT sheet as a document. 3) the base for the nodes of the output tree. Imagine a use case where you are using relative URIs with the following bases : ./xml/ for the document ./xsl/ for the stylesheet ./doc/out/ for the output tree If we allow relative URIs references and for absolutization, to map a namespace defined as xmlns:ns="foo" in the input document, you would have to declare it as xmlns:ns="../xml/foo" in the stylesheet (if I understand correctly the debate about "absolutization") and the transformation would have to generate xmlns:ns="../../xml/foo" in the output tree to keep a relative reference to the same namespace. It can probably be done properly and this kind of core XBase mechanism may even be useful to implement XBase extensions for XSLT to be used with XLink, XPointer and possibly XInclude. It requires, though, a modification of XSLT. Eric -- ------------------------------------------------------------------------ Eric van der Vlist Dyomedea http://dyomedea.com http://xmlfr.org http://4xt.org http://ducotede.com ------------------------------------------------------------------------
Received on Wednesday, 24 May 2000 14:21:46 UTC