Looking through the PMP/PWG notes, some issues are not quite clear.
1. Evolution of the Interfaces Group.
The resolution is not identified. Presumably the intent is to
reference RFC1213. What about the use of the extended interface types
in RFC1573?
2. Serial Number and Administratively Assigned Name.
Can we put these in a separate group so that they can be conditionally
mandatory? I guess we can end the fiction that the MIB is for all
printers. Obviously it is only for higher end network printers. The
argument that components have serial numbers is not valid, since it is
only high-end optional components that do. Obviously, I would have
voted against this.
3. Localization
I know I may not always be clear, but I cannot figure out how one of
my emails was taken to be a comment on localization. Hope it was a
good discussion.
4. Channel Type
As Jay pointed out, the e-mail requested changes between port 9100 and
AppSocket do not reflect what was originally submitted by Adobe or HP.
The changes and the origin of these changes should be identified.
5. Standard Printer Adapter Project
Don has not yet started the discussion but, as a supplier of NIC cards
to OEM's I think it a poor idea.
a. DPI has a 'standard' NIC design (DPO), supported by many
controller design houses and which can be second sourced.
Peerless/Emulex have the PSIO which, I understand, is also offered as
a standard. Several other companies produce 1284-type devices.
b. Large companies such as HP and Lexmark have their own
implementations, and are typically loath to share them.
c. The demands upon network printers (such as the Printer MIB) will
tend to obsolete the idea of networking a personal printer with a
card. The trend will be to on-controller network support with or
without network coprocessors.
Therefore, I suggest that
a: the niche market is already covered and has several different
implementations and suppliers to choose from.
b: any new standard would, presumably not match anyone's existing
implementation. Therefore, it would create yet another configuration
to be implemented. And being designed by committee, it would probably
be suboptimal.
c: however, there may be some rational in attempting to define a
print/MFP API for embedded network interface implementations.
Bill Wagner, DPI