ietf-mxcomp
[Top] [All Lists]

RE: Comments on draft-ietf-marid-core-01 xml use

2004-06-04 14:38:20

I agree that you can't have a registration for "every urn" 
under MARID.
The document says that the urns have the form:

urn:ietf:params:xml:<class>:<id>

I am still unable to see why there is an issue here.

The idea I had when I originally proposed using URIs as an OID substitute
was to open up the PICS protocol so that anyone with a contribution to make
could do so.

There is only one feature of a URN that is significant with respect to XML
schema, is the identifier unique. urn:ietf:params:xml:ns:marid-1 works fine
as an ID for MARID purposes. If the schema is revised the next namespace ID
could be urn:ietf:params:xml:ns:marid-2 or from a protocol perspective
gopher:info.cern.ch/pics/bugsbunny.jpg would work just as well.

There is a heck of a lot of confusion in the naming space. The original idea
of URNs was that the semantics of a URN are NOT defined by the resolution
process. That got fuzzed up a lot en route and a lot of broken ideas got
thrown in. But as far as XML schema goes there is no requirement for the URI
to support resolution. If that had been a requirement the name would have
been specified as a URL.


Is this the first time that the IETF has registered an XML schema??? Can't
we just do what the last guys did?

The way I would have organized the IETF namespace would be to partition off
a separate area for each WG since the WG names are guaranteed to be unique.
That way the system would work without the need for IANA intervention - and
since that now costs over half a million dollars a year that would have
seemed like a good idea to me.

urn:ietf:params:marid:1.xsd

The above would work for me. For XKMS I simply used the RFC URNs as a
substitute protocol identifier, it works.

        Phill