Repeat: XML URN namespace

[Jon Bosak <bosak@boethius.eng.sun.com>]

I sent this message out Friday, but Una didn't get a copy.
I'm not sure what happened.

Jon

------- Start of forwarded message -------

Date: Fri, 21 Apr 2000 12:29:28 -0700 (PDT)
From: Jon Bosak <bosak@boethius.eng.sun.com>
To: xmlorg@lists.oasis-open.org
cc: michael@bailey.dscga.com, rdaniel@metacode.com
Subject: XML URN namespace

I have been in correspondence with Michael Mealling and Ron Daniel
regarding the procedure for registering a URN namespace identifier
with IANA.  Michael has very generously offered to help us out
with this and given me permission to forward the message below to
this list, which includes people working on the first
implementation of the xml.org registry/repository.  Michael and
Ron, you should be able to post to this list, but I don't think
you can subscribe to it (or would want to), so people will have to
remember to include your addresses in messages concerning this
subject.

Jon

- ------- Start of forwarded message -------
 Date: Wed, 19 Apr 2000 16:33:14 -0400
 From: Michael Mealling <michael@bailey.dscga.com>
 To: Jon Bosak <bosak@boethius.eng.sun.com>
 Cc: michaelm@netsol.com, rdaniel@metacode.com
 Subject: Re: XML URN namespace....

 On Wed, Apr 19, 2000 at 01:00:20PM -0700, Jon Bosak wrote:
 > | Yes. In order to ask for a 'formal' NID (which is a NID that you
 > | request instead of just an assigned number) you must define the
 > | namespace in an at least an informational RFC (RFC status doesn't
 > | need to be complete in order to make the request). You are allowed
 > | to have a proprietary or opaque structure if needed but in your
 > | case I would advise against it.  It sounds like your asking if you
 > | can get the NID now and define it latter?
 > 
 > Sort of.  I think what we want to do is to define a very simple
 > syntax now (probably automatically generated for each entry from
 > the 11179 metadata provided by the submitter) with hooks for
 > alternative syntaxes as we get some experience with how submitters
 > use the repository.  So, for example, if we were granted the NID
 > "xml", then our initial implementation might automatically assign
 > a URN of the form
 > 
 >    urn:xml:auto:[generated identifier here]
 > 
 > (I'm making this up, of course), and then later we might add
 > 
 >    urn:xml:fpi:[SGML FPI assigned by the submitter here]
 > 
 > and add other schemes as we go.  Each of these schemes would have
 > to be formally defined, of course, but I'm not sure how to handle
 > this kind of controlled extensibility in the context of RFCs.

 Yes. This is very valid and has already been done with the IETF
 documents namespace. it looks like this:
 urn;ietf:rfc:####
 urn:ietf:bcp:####
 etc...

 And the registering document specifies that others may be added
 latter and specifies how that's done (either by updating the
 RFC, by referencing some locally defined file, or whatever makes sense
 for you).

 > | Let me know if you need some help with what you can and can't do
 > | (you can pretty much do anything as long as its unique,
 > | non-reassigned, and can be encoded in a UTF-8 string of
 > | characters). A good first step would be a list of the things you
 > | want to name. Terry made it sound like you not only wanted to name
 > | schemas but also schema registering organizations. Are there any
 > | other classes you need names for?
 >  
 > We need to name schemas, DTDs, stylesheets, descriptive documents,
 > and all the other items associated with a formal submission to the
 > registry/repository.

 One key to remember is that what really determines if there are
 differences in the syntax of the name is whether or not its required
 by the assignment or maintenance of the namespace. I.e. how the
 name is used is not part of the name itself. So if schemas and DTDs
 have one management style while descriptive documents have another
 then those two should be different. 

 Lets say for example that descriptive documents have to be approved (much
 like the RFC series). Then the number or name assigned to that document
 then would make sense as the actual URN as well.

 On first (and fairly ignorant glance) I would say that schemas and DTDs
 make sense as one scheme which would be something like a unique, owner
 requesting string. Descriptive documents would be another and would
 depend really on their maintenence regime. One major one I can see
 is one for the principal organizations and individuals involved.
 In the domain-name registration process we have something like this
 called a handle (mines mm2136). As an interesting suggestion we (NSI)
 is creating a URN namespace for people and organizations.

 I think it might help me to help you if I could find a copy of 11179.
 Is it avaialble or do I need to pay ISO?

 - -MM

 - -- 
 - --------------------------------------------------------------------------------
 Michael Mealling	|      Vote Libertarian!       | www.rwhois.net/michael
 Sr. Research Engineer   |   www.ga.lp.org/gwinnett     | ICQ#:         14198821
 Network Solutions	|          www.lp.org          |  michaelm@netsol.com
- ------- End of forwarded message -------
------- End of forwarded message -------