Re: Catalog requirements

[Terry Allen <tallen@sonic.net>]

| Catalog Requirements
| 
| Norman Walsh
| 
| Staff Engineer
| Sun Microsystems, XML Technology Center
| $Id: req.xml,v 1.3 2000/11/10 16:48:07 ndw Exp $
| 
| 10 Nov 2000
| 
|  Revision History
|  Revision 0.1                     28 Aug 2000                    ndw
|  First draft
|   ------------------------------------------------------------------------
| 
| This document lays out the requirements for developing an XML-based version
| of [TR 9401:1997].
| 
|    * The catalog must be expressed in XML ([XML]) with a specified
|      namespace ([XML Names]).

Further, there should be a DTD for the catalogue.

|    * The catalog must provide the ability to map public identifiers to
|      URIs.

Both to URLs and to URNs (passim).

|    * The catalog must provide the ability to map system identifiers to
|      URIs.
| 
|    * When both a public identifier and a system identifier are present, the
|      author of the catalog must be able to express which should be used
|      preferentially.

Are present in what?  I think you want to phrase this differently.

|    * The catalog must support xml:base ([XML Base]).
| 
|    * The catalog must provide the ability to delegate mapping of classes of
|      public identifiers and URNs to alternate catalogs.

to *other* catalogues.

|    * The catalog must provide the ability to map notation names to URIs.
| 
|    * The catalog must provide the ability to map entity names to URIs.

Re John Cowan's objection:  I've forgotten (already!) whether we are
agreed not to support SGML features not present in XML, but it sure
would be useful to do so.

|    * The catalog must provide the ability to map URIs to alternate URIs
|      (for example, namespace names).

to *other* URIs.

|    * Where the functionality of the XML catalogs developed by this
|      Technical Committee is equivalent to functionality provided by
|      TR9401:1997 Catalogs, the semantics must be identical.
| 
|    * The catalog must provide an extension mechanism for experimental
|      additions to the catalog.


regards, Terry