- From: Martin Bryan <mtbryan@sgml.u-net.com>
- Date: Sat, 28 Jun 1997 13:24:09 +0100
- To: w3c-sgml-wg@w3.org
At 12:03 27/6/97 -0700, Tim Bray wrote: >No, we don't want a lexical model. We want CDATA, but unique. Which >is not too much for XML. But it's violently non-8879-conformant. -T. There is nothing to stop XML from saying that the values for any CDATA attribute called id (but not ID as this is a case sensitive system:-)) in an XML document must have a unique name. The reason for IDs being Name and therefore case sensitive is that for SGML all names are case sensitive. This allows you to put <p> and <P> and <p ID="x"> and <P id="X"> without fear of creating a clash because you were in the wrong shift at the time you entered it. For most users this is essential. If you ever try to use a system with NAMECASE GENERAL NO in its SGML decalaration you will soon experience the problems it creates. WG8 has discussed the problem with ID shift before, and will probably offer a switch that allows different rules to be applied to names within attribute values (including the use of digits in NAMES as well as NMTOKENS, which is the currently valid way of allowing names that start with numbers). It will probably have to add another switch to provide optionality for ID naming [something along the linees of IDTYPE=(NAME|NMTOKEN|CDATA)]. I would bet that after a few weeks of use the CDATA switch will be abandoned in the face of frantic user complaints. For the time being I would suggest that XML simply removes ID as a class of names and recommends that users use of a CDATA called id, and applies an XML-specific semantic requiring uniqueness of values for this attribute within a document. (And within any elements users may choose to add to the document through SHOW=EMBED.) Such an approach would not be violently non-8879-conformant. ---- Martin Bryan, 29 Oldbury Orchard, Churchdown, Glos. GL3 2PU, UK Phone/Fax: +44 1452 714029 E-mail: mtbryan@sgml.u-net.com For details about The SGML Centre contact http://www.sgml.u-net.com/ For details about the Open Information Interchange initiative contact http://www.echo.lu/oii/home.html
Received on Saturday, 28 June 1997 08:25:04 UTC