> My understanding is that Keith is trying to dictate that IPP CANNOT USE
> "http" - full stop.
No, that's not quite what I meant.
What I am "trying to dictate" is that IPP traffic must be easily
distinguishable from HTTP traffic, so that it can be filtered (or not)
according to a site's security policy. My suggestion to use a
different default port was an attempt to acheive this, with the fewest
possible changes to the current IPP protocol.
IETF has traditionally used well-known port numbers to distinguish
between different services. To follow this pattern, IPP should not
use port 80 as a default, because that port is reserved to HTTP.
And in my mind this pretty much implies that a new "ipp" URI prefix is
needed to refer to printers and print jobs so that the port number
doesn't have to be explicitly specified. This doesn't necessarily
mean that "http" cannot also be used (and doing this might be useful
to tunnel through proxies that understand http: but not ipp:) but
sometimes it's a Bad Idea to have two ways to name the same thing.
(What happens if you make a request to an ipp: object? Will you get
back references to ipp: objects? or might they use the http: scheme?)
Note that while some changes might be necessary for IPP protocol
elements (using ipp: URLs instead of http: URLs) I would not expect
any changes to the HTTP layer itself.
Keith