- From: Peter Murray-Rust <Peter@ursus.demon.co.uk>
- Date: Mon, 21 Apr 1997 13:51:08 GMT
- To: w3c-sgml-wg@w3.org
I have been struggling with the precise implementation of XML-LINK 5.2 and 5.3 and while I think I have the intention of the draft, I find ambiguity. 5.2. Production [1] states that after a connector there may either be a Name or an Xpointer. Xpointer is defined [4] but there is no definition of Name. (I assume this to be the same definition as in XML-LANG, but the BNF namespaces are apparently separate (Element is defined in both and differs - unless 'element' and 'Element' are held to be different). The construction HREF="Foo#HERE" is therefore ambiguous since it is unclear whether HERE is an attribute of type ID or the reserved word HERE (similarly for ROOT). 5.3 second bullet. 'beginning and end of a span' ... 'FROM and TO'. This is unclear to anyone not from the TEI community (which includes me :-). In 5.3 below it talks about 'all the text from the location'. I am unclear what to do with the following (please imagine there are no newlines and inter-tag whitespace so magic #PCDATA isn't involved): <COMPANY ID="FOO"> <DIVISION ID="MIS"> <EMPLOYEE ID="DON"> <EMPLOYEE ID="SUE"> </DIVISION> </COMPANY> What is the result of the valid query: #ID(FOO)..ID(DON) It might be: (a) a null string since there is no _text_ enclosed (b) invalid since it could be seen as not well-formed (c) equivalent to: <COMPANY ID="FOO"> <DIVISION ID="MIS"> <EMPLOYEE ID="DON"> </DIVISION> </COMPANY> [A general comment is that this section is written as if the documents are primarily text with a few embedded tags. It's quite meaningful to create XML documents with no #PCDATA at all.] 5.3.1 ROOT is twice referred to as 'the default' and this fooled me into thinking it could be omitted. I have read this carefully and it is not a default since it has the same status as HERE and ID. The sentence 'the presence of the ROOT keyword has no effect...' seems to be wrong. If the ROOT keyword is omitted I hope an error would result :-) P. -- Peter Murray-Rust, domestic net connection Virtual School of Molecular Sciences http://www.vsms.nottingham.ac.uk/
Received on Monday, 21 April 1997 09:22:04 UTC