I included the comments that I sent out on Friday about empty values,
so that all Job, Job Template, and Printer attributes require at least
one value.
I did not include any of Bob's and Carl-Uno's comments. Scott will do
that this Sunday.
1. I did suggest a Printer attribute to contain the URLs of the associated
Job Templates with a qualifier that says which one is the default for the
printer and have included the text. I also added more description about how
Job Templates work and indicated an issue about naming Printers in the
Directory that do and don't imply Job Templates. And I added a Job Template
URL as a third type of URL that a client can supply in a Get Attributes
operation in order to get the defaults as is mentioned in two places
in the draft.
2. Since we've agreed to keep job-retention-period, I add it has an input
parameter of the CancelJob operation (as in ISO DPA), so that the user
can decide whether the keep or not keep the job when he cancels it.
3. I added Requested Attributes to Print and Get Jobs operations, so
that they are the same as Get Attributes operation.
4. I indicated that end-user operations is not in V1.0, to leave the door
open for operator operations being added to a later version of IPP.
5. I added the definitions of "shall", "should", "may" and "need not" and
started to use that conformance languages in the specification.
6. I removed redundant specification of values for the attributes.
The specication for most values is in the Job attribute. A few directory
attributes do not have corresponding job attributes, such as color-supported,
so the directory attribute values are also type2Enum.
7. I did not touch the HTTP stuff. One issue we have is the case of
attributes. Roger's flows indicate the first letter being capitalized,
while all our attribute specifications start with a lower case letter.
To get the document out, it may be easier to change the flows, rather than
all of the attributes and their references.