Re FPIs for RFCs

[David G. Durand]

XML-dev is not dependable for me. I think this posting was lost.

Summary:

FPIs cannot be assigned in someone else's namespace, the "owner" identifies
a naming authority, not an intellectual property owner. Assertion supported
with verbiage quoted from relevant ISO standards.

At 4:11 AM -0000 12/2/97, Rick Jelliffe wrote:
>
>> From: Terry Allen <tallen at sonic.net>
>
>> You are asserting an ownership right you cannot back up.  That's dangerous
>> for one's legal health.  Referring to something by using its URL is one
>> thing, but using that URL to create a name that lies in someone else's
>> name space is another matter entirely.
>
>I have emailed Internic to find out their views.  However, I do not
>believe that an FPI is property.

I'm sorry this is flat out wrong. ISO 9070 is very clear on the subject,
and I quote:

"3.10 Owner name: the portion of a public identifier that names its owner.
NOTES
.... 13 The owner of a public identifier is not necessarily the owner of
the object it identifies"

and from the introduction:
"... and an 'owner name', which identifies the originator of the public
identifier"

The whole point of owners (I can't quote the showe standard, unfortunately)
is to create domains of administration for namespaces, and sub-namespaces.
This just can't work if I'm allowed to make names in _your namespace_
without your permission, just because I'm citing your work. So either there
is a wording goof in 8879, or 9070 is making

> I believe it is common and accepted
>practise to create FPIs for published material using ISBN

I've never heard of this practice, and since it leads to a chaotic and
broken system of public identifiers, we should stamp it out to the extent
that it has been accepted.

> and that
>the IDN can be used in exactly the same way.

IDN is not in 9070 rev 2, and thus is not suitable _de jure_; it is also
unsuitable _de facto_, since domain names can be reused by different
organizations. Unless Internet policies and 9070 have both changed, I think
this is also wrong.

>ISO 8879 says 4.223 owner identifier "The portion of a public identifier
>that identies the owner or originator of public text".

That definition conflicts with the 9070 definition, but in the context of
the public entity sets in SGML, the confusion of identifier owner and data
owner is understandable. When 8879 was written, the notion of needing to
assign persistent names to _other people's_ computer readable documents was
not foremost in anyone's mind.

>I read that to mean that it would actually be wrong for me to use myself
>in the owner field.  The owner means the owner (or originator) of the
>public text, not the originator of the FPI.

This interpretation is explicitly contrary to 9070, though, and 9070 is
more recent, normatively cited by 8879, and edited by the same editor; so I
am inclined to prefer the 9070 reading. Also, the name assignment protocol
you suggest fails to achieve the list of goals in 9070 for the whole public
identifier standard, since it fails to provide a set of rules that will
guarantee that no colficting PUBLIC identifiers are ever assigned. The 9070
interpretation does provide such rules, based on hierarchical name
assignment.

  -- David

_________________________________________
David Durand              dgd at cs.bu.edu  \  david at dynamicDiagrams.com
Boston University Computer Science        \  Sr. Analyst
http://www.cs.bu.edu/students/grads/dgd/   \  Dynamic Diagrams
--------------------------------------------\  http://www.dynamicDiagrams.com/
MAPA: mapping for the WWW                    \__________________________



xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)