Meeting notes from the
informal UPDF meeting in
Portland, June 26th/27th 2002.
Attendees
Jim Sommer, Granite Systems
Mark Hamzy, IBM
Norbert Schade
Although there was no official UPDF meeting scheduled for Portland, Mark Hamzy, Jim Sommer and I joined parts of the PSI meeting and the full-day plenary to represent our interests.
However we took the time to meet on Wednesday afternoon (6 hours) and Thursday evening (two hours) to discuss some open issues we could not resolve in emails and phone calls.
These were the major issues:
1. ICC profiles
We will add an element References under 'PrintCapabilities' like 'Features' and 'Objects.
The only element underneath will be 'ICC_Profiles'. That way we can list as many profiles as we want and can select them with dependencies like features and objects.
Expect to hear more the next days.
2. Hardware margins
We have added an attribute 'Orientation' to 'HardwareMargins' and could now save an extra margins element for landscape. It's easier to implement it this way on the host.
3. Dependencies
We shortened the term 'Interdependencies' to 'Dependencies'.
The structure was changed as well. We do not support the definitions of 'OR' conditions any more. It is too complex to write a parser, which would understand all possible varieties of conditions. And we want to encourage people to read the xml on-the-fly.
Now it is necessary to split previous 'OR' conditions into separate conditions. On the other hand we have a simpler structure. It is still possible to combine as many features as you want in one condition.
And the complete action section is more or less unchanged.
The attribute 'Relation' is now constrained to 'equal' and 'not-equal', as we did not find any sample where 'less' and 'larger' really was needed.
More details to come for the interested expert.
4. Classifying identifyers
In general we are intending to consistently work with Predefined_ID/Proprietary_ID attributes wherever possible. You will see a number of related changes. This eases the host implementation as well.
5. Syntax problems
We were discussing some global syntax problems for some time, but we could solve all of them.
One major change to accomplish that is to allow IHV to add their own namespace to UPDF if needed.
The cases where useful of needed will be rare, but it provides for a consistent syntax all over the place when it comes to command sequences and parameters referenced within them.
See samples coming up.
6. Common PWG Semantic Model
The PWG declared the IPP concept the base for the common Semantic Model and defined a set of schemas to document that.
We will check, how close UPDF already is to these structures and discuss open issues with them.
That will be one of our main activities the next weeks.
7. UPDF as common base for device capabilities under IBM Linux solutions.
There is some discussion going on in a Linux forum, where the PWG and IBM are actively contributing.
One option is to take UPDF as their common format to store driver capabilities by using it.
This is a very interesting approach and proves that the UPDF standard is on an excellent level.
This is another activity we will persue in the near future.
8. Print Services
As many may know there is a new group in the PWG community called Print Services.
They are checking UPDF and probably may use some of the functionality to accomplish some of their targets.
Another field where the hard work we did over the last years will pay out.
We are supporting them as much as we can, as we consider that activity very close to a driver.
9. We invite all IHV to start some UPDF implementations, as we have the host implementations under MS Windows and IBM Linux working to quite a good level. Now it's the time to get device descriptions checked.
So start writing them and let us know, if you need help.
We will send another note through the PWG reflector to get the message around.
You see we could cover a lot of bases in two very effective meetings. We do not plan to meet in Santa Fe, but most likely we will meet in New Orleans. In the meantime we will continue with the host implementation and finish the UPDF design for level 1 section by section.
We agreed that we will not meet in Santa Fe, but will likely meet in New Orleans.
Norbert Schade