- From: Paul Prescod <papresco@calum.csclub.uwaterloo.ca>
- Date: Fri, 16 May 1997 14:25:31 -0400 (EDT)
- To: robin@acadcomp.sil.org (Robin Cover)
- Cc: w3c-sgml-wg@w3.org, robin@acadcomp.sil.org
Robin, I think that there is theoetical benefit to make philosophic distinctions between objects and attributes, but the practical benefit is probably quite small. I suspect that this is the reason that SGML has been dodging this issue for so long. The practical benefit proposed (backwards compatibility of DTDs) is not really there in many (most?) cases. A few other misc. notes: #1. The grove model of attributed graphs does not prohibit "properties" from having trees as values. In fact a nodes children are always specified through a property. That may or may not be relevant in this discussion. #2. You can simulate tree-valued attributes without implying containment through ID/IDREF or even ENTITY attributes. Paul Prescod
Received on Friday, 16 May 1997 14:25:42 UTC