Great summary. I agree.
>>> "Scott Lawrence" <lawrence at agranat.com> 05/30 12:16 PM >>>
>>>>> "LM" == Larry Masinter <masinter at parc.xerox.com> writes:
LM> The difficult issue in print-by-reference is how the sender
LM> supplies the credentials if any at all are required. There
LM> is no single standard for identity in Internet protocols, and
LM> there's no standard encapsulation of credentials.
Agreed; which, I presume, is why Roger K deBry [RKD] suggested that
only 'publicly available' [a question-begging term, IMHO] documents
could be printed by reference:
RKD> The two conditions required are that the document be in a
RKD> standard printable format and that the document be publicaly
RKD> available.
My answer is to say that no credentials of the user are employed by
the printer to obtain the referenced document; instead, whatever
credentials of its own the printer has are used. This degenerates
to Rogers' model when the printer has no credentials of any kind
(but note that the document repository might choose to consider the
source IP address or domain name of the printer to be a form of weak
but acceptable credential). You might need an extra status or two
to communicate the inability of the printer to access the referenced
document, but I suspect you need that no matter how you do this.
--
Scott Lawrence EmWeb Embedded Server <lawrence at agranat.com>
Agranat Systems, Inc. Engineering http://www.agranat.com/