ietf-mxcomp
[Top] [All Lists]

RE: suggested new RRtype experiment

2004-05-21 17:43:08

On Fri, 2004-05-21 at 17:19, Bob Atkinson wrote:
If this schema is fixed as part of the specifications, then adding any
reference other than the registered version (document type) is not
needed nor desired.  This http link is offering a mechanism for ad hoc
modifications or extensions.  It is not, by any means, for the
programmer's convenience.

You're quite wrong about this. 

The schema used in the URI isn't important. The only criteria used to
choose the existing Caller ID namespace was that it be short.

But the fact that there is a URI in the document, which names the XML
namespace in which the elements etc are found, is quite critical.

Not having a URI means there MUST be a predefined definition within a
specification referenced by a much simpler and smaller tag.  Only the
tag is critical.  Having a URI is highly problematic if to lock down
definitions.  IANA already has a registry for these definitions.  If
these definitions change, the tag changes per the specification that
provides the changes.  As these specifications are to be used to compile
a parser, it is important these specifications not change in an ad hoc
fashion or the model of operation changes in a dramatically bad way.

-Doug