- From: Rick JELLIFFE <ricko@gate.sinica.edu.tw>
- Date: Fri, 18 Feb 2000 03:26:45 +0800
- To: www-xml-linking-comments@w3.org
I still cannot see why XBase is needed. 1) For XML, we can use a entity reference. XML Schemas are not replacing DTDs, because they operate on the infoset level not on the tokens: hence XML Schemas do take over from DTDs the entity declaration mechanism or the ability to declare a default namespace using a (e.g. #FIXED) attribute in a DTD. So XBase cannot be justified by any immanent disappearance of DTDs. So XBase duplicates an existing mechanism for XML documents, with the addition of some scoping of uncertain value. 2) We already have an element defined with this functionality: <html:base> If XLink does not need to create a new mechanism for dealing with XHTML, it should merely recognise <html:base>. So a new attribute is not required for supporting html. 3) For XSL transformations, bases can just as easily be held in variables and generated in full in the output tree. XBase looks like a uneeded complication, at a time when many people want "simplicity". I am particularly concerned that there seems to be no Requirements Document for it, since that would help me judge what on earth it is for. I would hope there is no chance of anything going to Last Call without a Requirements Document first. Rick Jelliffe Academia Sinica Computing Centre
Received on Thursday, 17 February 2000 14:24:16 UTC