- From: <petsa@us.ibm.com>
- Date: Fri, 9 Apr 1999 09:54:28 -0400
- To: rbourret@ito.tu-darmstadt.de, www-xml-schema-comments@w3.org
- cc: murray@muzmo.com, C M Sperberg-McQueen <cmsmcq@acm.org>, connolly@w3.org
Ron: This is in reply to your comments on the XML Schema requirements document sent on February 26. I apologize for not getting back to you earlier. This is not sufficient excuse but we have been busy with design work and, you will be happy to know, it is going well. You raised the following specific points in your note. 1. Clarify relationship between schemas and DTDs. The requirements document says that schemas shall be "more expressive than DTDs". By this we mean that schemas will be a superset of DTDs in generative power. As for readability, schemas will have XML syntax and I judge that to be more readable than DTDs but that's a personal preference. 2. Reuse mechanism simpler than inheritance Several people have asked for mechanisms other than inheritance that would further reuse. This is currently under investigation by the design team. 3. Remove entity definitions from schema language. This, too, is a design issue and is being considered by the design team. 4. Datatype requirement 2 may be too hard. There are several issues re. mapping from databases to XML. We are attacking some of them. For example, we are defining datatypes that enable simple mapping from database values to XML. The issue of keys is a thorny one. I personally believe that the ID/IDREF mechanism in XML is not adequate as a general reference mechanism and would like to see something better but I am not sure we will step up to representing primary keys and foriegn keys in the first draft. Thank you for your comments. All the best, Ashok Malhotra
Received on Friday, 9 April 1999 09:54:50 UTC