New Internet-Draft Request

New Internet-Draft Request

New Internet-Draft Request

Robert Herriot robert.herriot at Eng.Sun.COM
Fri Nov 22 22:42:24 EST 1996


Great job Scott in pulling this all together. My comments are below
from reading the changes. I may have more comments after I read the
text that didn't have revision marks.


Lines number are with respect to ipp93rev.doc document in MS Word with Laser
Writer Plus Printer. I hope they align with your numbers.  


By the way mail servers will be down Sunday due to a planned power
outage. So if my mail bounces, please try again on Monday.


line 468:  I am not sure that I agree with your change to "zero or more Job
Template objects" associated with a Printer. I think that we should leave 
it as an issue as to whether a Printer accessible via the print operation 
can have no job template associated with it.  


line 527: Is 380 now a registered port with IANA for HTTP printing?


line 544: As someone suggested in email. The name service might have objective
values and the name search software might take either objective or subjective
values from a client.  Somewhere else there could be a mapping.  
The main point that I think should be added here that there are a wide 
range of possible solutions.  It may not be necessary for all name service 
entries to have both subjective and objective entries.


line 593: In addition to four color (CMYK), there is 3 color (CMY).  The 
names CMYK and CMY should probably be in the name in case some "printer" uses 
RGB for color.


line 616: Change "Is is" to "It is", and change "recommended" to "required";
Otherwise, the syntax is very useful.


line 692: Delete "binding" from "long-binding-edge"


lines 750 and 757 should be indented further.


line 793: There is no mention of what the HTTP method should be for the four
operations presented in this section. We have talked about POST for Print and
CancelJob and GET for GetAttributes and GetJobs.


lines 850: Although I think most of this HTTP structure is right, I think
that we need to discuss it further. In particular I wonder whether
documents should be formal HTTP Entity-Bodies and have Content-types,
such as Application/PostScript or text/plain. We just need to recognize
that this area needs further discussion even whether it is flagged as an
issue in the document.


line 878 and 887: the syntax can be shorter as:
        Attribute Value = #Value


line 904: does the last item of the table imply that URL is an attribute
  and not a document-content.


line 918: There should be some mention about authorization required for 
   Cancel Job.


line 934: "Ignored by the Server" is this in the context of null attributes.
  If so, what is there to ignore. When the client supplies no attributes,
  what does the server return?


line 941: Isn't the Printer or Job URL implicit in the HTTP operation?	


line 951: we may want a few more default attributes to be return, such
   as the job-identifer and the position in the queue if not implicit.


line 957: the semantics in the above paragraph suggest that the client
   can specify attributes, but the table entry does have this parameter.


line 963: same as line 951. Job identifier seem especially important because
  this operation may be the only way a user can figure out the name to
  use to cancel a job.


line 1225: correct the definition of "#" syntax to be that of rfc 822
  per email discussions, i.e. zero or more item separated by a ",".


line 1370: missing header and header runs into paragraph. There 
   should be an issue as to whether this should be a boolean value.


line 1469: Tom raised the issue of deleting retention time and then
  reversed himself after I and Sato objected. This may not be an issue
  any longer.


line 1541: I am not sure than these words on embellish are what we
   all agree to. Perhaps, we should leave some alternatives as
   issues. At the end of the discussion this afternoon, Tom and I
   agreed that 'none' mean no number-up and '1' might be 1 page/impression
   with default number-up embellishment.  But let's leave this as
   an issue.


lines 1555, 1556: remove "binding" from "binding-edge".



More information about the Ipp mailing list