- From: Martin Bryan <mtbryan@sgml.u-net.com>
- Date: Sat, 24 May 1997 10:47:33 +0100
- To: w3c-sgml-wg@w3.org
At 17:45 23/5/97 GMT, Christopher R. Maden wrote: >I'm not saying that colons are the way to go - but I think that >overloading CONCUR in a way incompatible with ISO 8879:1986 is *not* >the way to go. The question is not one of overloading - its a question of how can we integrate all the name space proposals into a single whole when SGML is extended to take care of name spaces. SGML currently has two ways of identifying name spaces - DTD/LPD names that are used to qualify the names of elements or define which set of attribute definitions should be applied, and entity names that are used to distinguish the name spaces of entities used as subdocs. We already have a way of identifying IDs associated with subdoc entities (entity+id) and a way of distinguishing which IDs belong to which DTDs (the name group for the element for which the ID is defined). What we need to ensure is that any mechanism to used by XML can be generalized into a rule that does not clash with, but extends, these existing name spaces. I personally would like to see a single mechanism that would allow us to identify things based on a combination of entity-name+doctype-name+id. The problem is how to associate an instance specfic name, such as the entity name assigned to each subdoc, with fragments associated with a particular doctype name so that IDs from different instances conforming to a single DTD can be embedded in a single document without a namespace clash. This is really difficult in SGML, where control of external data is done using named entities - it is almost impossible in a DTD-less world such as XML. ---- Martin Bryan, The SGML Centre, Churchdown, Glos. GL3 2PU, UK Phone/Fax: +44 1452 714029 WWW home page: http://www.sgml.u-net.com/
Received on Saturday, 24 May 1997 05:47:27 UTC