Eric,
I like the list of uses of the Resource object in your second paragraph:
ICC Profiles, etc. Your list shows exactly why such a Resource object would
be useful.
As to the answer to your question about images and how they are positioned,
the PDL takes care of that. The PDL has some mechanism to reference files.
Postscript "include" I believe it is called. There is also an "Open
Prepress Interchange (OPI) specification for PostScript comments that make
references to high resolution images.
Search the Adobe partners web site for OPI to locate these specifications
(i.e. at http://partners.adobe.com <http://partners.adobe.com> )
So images are just another example of a resource that could be down loaded
into a printer so that various documents could make reference to them.
Yes, we should look at what the PPML folks are doing for PODi as well. But
this IPP Resource object and operations proposal is just to provide means to
load, query, and get such images, not extending PDLs to make reference to
them. So if PPML (like PostScript) has ways to make external references to
other PPML files or other objects, then PPML would benefit from having the
resource object and its operations defined as well.
Tom
-----Original Message-----
From: Eric Random [mailto:ERandom at peerless.com]
Sent: Wednesday, May 10, 2000 18:13
To: hastings at cp10.es.xerox.com
Cc: ipp at pwg.org
Subject: Re: IPP> Sketch of an idea: Define a Resource container object for
Print Driver Extension and other sub-types
Tom -
While the Resource object seems like a potentially useful discussion topic,
I'm not sure that I see the relationship between several of the proposed
sub-types and IPP. For example, how would a client use IPP to position an
image stored in a library on the device correctly in a printed document?
Aren't the PPML folks over at PODI working on a document schema based on XML
to solve the general class of variable data printing problems? I know I
would benefit from a more complete discussion of where you are headed with
this, since at first glance it seems contrary to the principle of separating
job submission, management, and monitoring protocol (IPP) from document
content.
I guess my real question is "Are you trying to define a general purpose
printer resource selection mechanism by extending IPP syntax?" If you are,
I'd suggest that you also consider additional resources typically associated
with color printers, such as ICC profiles, Color Rendering Dictionaries, and
other by-products of print engine characterization, including (but not
limited to) halftone screens (possibly implemented as threshold arrays) and
grey balance transforms (possibly implemented as 1-D lookup tables). Such
resources can be downloaded into a printer, selected from a pre-existing
store in the printer, or even uploaded from the printer for use by a host
based RIP (there is existing standard PDL syntax to accomplish the first two
operations, but I don't know of any standard method for accomplishing the
latter).
Eric Random
Core Technology
Peerless Systems
(310) 297-3251
(310) 727-5715 - Fax
erandom at peerless.com <mailto:erandom at peerless.com>
>>> "Hastings, Tom N" <hastings at cp10.es.xerox.com> 05/09/00 11:16AM >>>
We have discussed several times on the IPP telecons about adding a new
Resource object. This mail note is a sketch of the idea to see if there is
support for developing an IPP spec along these lines. The reason to being
this idea up now is because of the discussion of the Print Driver Extension.
The Resource object approach would allow a way to install print drivers on
the Printer and to Get Print Drivers from the Printer to install on a
client. It would allow a Printer to have more than one print driver, say
one for each of a number of different OS platforms and/or PDLs.
Should the Resource object be an agenda topic for the New York City IPP WG
Meeting, 5/17-18?
Here is the sketch of the Resource object idea and its operations:
The Resource object is a generic container object for a number of sub-typed
objects. The Resource object would have a number of operations defined.
Each operation would include an operation attribute that identifies the
object sub-type in question. Then implementers could define new sub-types
easily without having to invent new operations.
There are some attributes common to all sub-types, such as the
"resource-name", "resource-sub-type", "resource-owner",
"resource-creation-date-time", etc., and a lot that are specific to a
particular sub-type. Some sub-types will also have opaque data associated
with each object instance (such as fonts, forms, images, and print drivers),
and others will only have attributes (such as media).
The initial list of sub-types include:
media - Just attributes which define the media characteristics. So the
client can query the media library and the administrator can add new media
and define their characteristics
fonts - attributes and the PDL data
forms - attributes and the PDL data
logos - attributes and the PDL data
images - attributes and the PDL data. high resolution images can be put
into a shared image library and called out from a number of documents
print drivers - attributes and the opaque data is the executable code
The operations would include:
Create-Resource - the client specifies the resource type and a user-friendly
name for it that is used in all subsequent operations, including Job
Submission operations (Print-Job, Create-Job, Send-Document). A lease time
is also requested (just like our Subscription objects) and the Printer
returns the lease time granted (which may be infinite meaning no need to
renew or finite, meaning that the client must renew the lease, else the
resource will be deleted).
Delete-Resource - delete the named resource of the indicated sub-type
Renew-Resource - renews the lease, in case it wasn't infinite
Get-Resource-Attributes - returns the requested (or all) attributes of a
specified resource type and instance.
Get-Resources - returns the requested (or all) attributes of a specified
resource type that matches the supplied filter criteria consisting of any of
the resource attributes. Here the filtering is more complex than we have
for Get-Jobs or Get-Subscriptions operations, since the client could filter
on any of the resource attribute values. For example, a client could
request all of the media that has a "media-size" equal to particular pair of
x and y dimensions and get back the other resource attributes of any matched
media instances.
Get-Resource-Data - returns the opaque data for those resources for which
the data is not copyrighted, etc.
These operations are very similar to the Subscription object operations that
the IPP WG has nearly approved for notification. The only differences are:
Subscription Ids are assigned by the Printer as numbers, while resource
object instances have user friendly names assigned by the creator. Both
have leases, but Subscription leases are expected to be much shorter
(minutes, hours) and handled by the software, not the user, while Resource
objects leases are much longer (days, weeks, months), though both can be
infinite depending on SA policy, and the user or administrator explicitly
creates the Resources. Also Resources are known to users and administrators
and are managed by them. Also Subscriptions don't have any complex filtering
in the query requests.
[We don't want to merge Subscriptions into Resources, just illustrate their
similarities]
It will also allow the administrator to set up policy on who can create
which kinds of resources and for how long.
Comments?
Thanks,
Tom