Is there any experimental scheme that we could use, so we don't have to muck
with the "ipp:" scheme?
If there is, we could just mention what that scheme is in the spec.
Or since this is a URI data type, is there some URN form that we could use
to give each document a unique identifiers? All we want is a unique
identifier across that one Printer. Its doesn't have to be unique across
all Printers, right?
We also need to agree as to over what time period the ID has to be unique?
If the Printer is powered off and comes back again, does it have to continue
to generate unique URIs that it never generated before? Or don't we have to
be that strict.
From: McDonald, Ira [mailto:imcdonald at sharplabs.com]
Sent: Tuesday, March 25, 2003 09:18
To: 'sm at pwg.org'; 'ps at pwg.org'
Subject: SM> document-id-uri semantics
The current "document-id-uri" in the IPP Document Object spec
clearly that states that it is an opaque identifier, and cannot
be used as the target of an operation (unlike the "job-uri" in
RFC 2911). And it gives an example of an 'ipp:' schemed URI.
I think I should write an Appendix to the Document Object that:
(1) extends the IPP URL Scheme (adopted last month by the IETF
for RFC publication as an IETF Proposed Standard) to apply
to Document objects as well,
(2) specifies the opaque identifier limitation,
(3) is normatively referenced by the "document-id-uri" attribute
definition in the main spec.
As we discussed on the PSI telecon this morning, PSI doesn't
care what the URL scheme is in the DocumentURI, just that it
be unique within a print server.
- Ira McDonald, co-editor of IPP URL Scheme
High North Inc