Bob,
I attach the current version of our current data types schema.
If I understand your vision properly, this should exactly demonstrate the
separation of attribute structures and the overall implementation - in our
case the UPDF device description schemas.
Depending on how close we'll come eventually, I could imagine taht we even
include a Semantic Model data type schema within the UPDF data type schema
and take the duplicates out of our current version.
But I don't want to be too optimistic, before I see a clear path of the
Semantic Model.
And there could be some other significant requirements specific to a UPDF
device description.
Regards
Norbert
(See attached file: UPDF Data Types.xsd)
"Harry Lewis"
<harryl@us.ib To: "TAYLOR,BOB (HP-Vancouver,ex1)" <robert_b_taylor@hp.com>
m.com> cc: Norbert Schade <norbertschade@attbi.com>, Print Services group <ps@pwg.org>, UPD group
Sent by: <upd@pwg.org>
owner-upd@pwg Subject: UPD> RE: PS> Semantic model: media handling
.org
07/17/2002
11:34 AM
I like Bob's analysis. I think the Semantic Model will be most useful
taking this approach.
----------------------------------------------
Harry Lewis
IBM Printing Systems
----------------------------------------------
"TAYLOR,BOB (HP-Vancouver,ex1)"
<robert_b_taylor@hp.com> To: Norbert Schade
Sent by: owner-ps@pwg.org <norbertschade@attbi.com>, Print
Services group <ps@pwg.org>
cc: UPD group
07/16/2002 09:13 PM <upd@pwg.org>
Subject: RE: PS>
Semantic model: media handling
Hi Norbert, all,
One of the things we (HP) have been suggesting for the semantic model is
the separation of the raw "attribute/element" definitions from the
structures/model that pull them together for a particular use. As you not,
UPDF has done this structuring in a different way than IPP - which is also
somewhat different than UPnP & PSI, etc. I'm not sure we want to try
codify any one structure as part of the core semantic model - these will
tend to vary by market segment and domain, and I'm not sure we can do this
one-size-fits-all. What we would like to see, though, is common definition
of the
core "attributes/elements" - this seems much more reusable across models &
domains. It does make sense, though, to publish some of the "common
models" as at least examples of structural models - IPP, UPDF, etc. are
likely candidates for this. This exposes some of useful constructs (such
as the composite feature you describe below) for reuse.
thanks,
bt
---------------------------------------------------
Bob Taylor
Senior Architect
IPG Strategic Technology Development
Hewlett-Packard Co.
mailto:robertt@vcd.hp.com
phone: 360.212.2625/T212.2625
fax: 208.730-5111
---------------------------------------------------
-----Original Message-----
From: Norbert Schade [mailto:norbertschade@attbi.com]
Sent: Monday, July 08, 2002 7:32 AM
To: Print Services group
Cc: UPD group
Subject: PS> Semantic model: media handling
I have problems to follow two different ways to specify media handling and
UPDF would have problems to support that.
I'm fine with the specification of single media attributes like size, type,
etc.
I agree that there should exist a media instance a level higher, which is a
media element with a number of media attributes.
The number of attributes can vary. In one sample it may be just size and
type, in another it may be something like the IPP media collection.
My point is that the attributes a media is described by may vary.
There should not be a predefined media collection in a common Semantic
Model representing one implementation.
Feel free to check the composite feature definition we have in UPDF. Open
the UPDF.xsd schema to do this and follow the path down to
PrintCapabilities.Features. The current sample description xml of an
imaginary LJ9000 has a 'Media' composite feature. We can compose any
number of features to a new feature, be it Media, Quality or anything else.
This is a very flexible structure and is expected to be used frequently. We
got very positive feedback once we finished it last year.
We'd appreciate if the Semantic Model does something down that path.
Otherwise the spec is ambiguous.
Another statement:
We've seen the current schema of the Semantic Model. We know there are a
number of ways to write schemas. The UPDF group made the experience that
working with attributes instead of assigning text to elements directly has
advantages. Validation is easier and we can define constraints (these are
really constraints and not dependencies) for attributes. You may think that
over.
Regards
Norbert Schade
69 Prescott Drive
North Chelmsford, MA 01863
978-251-1017
norbertschade@attbi.com
This archive was generated by hypermail 2b29 : Wed Jul 17 2002 - 16:06:04 EDT