> I would suggest that
> perhaps letting all of the Linux guest print via RSCS might be a
> reasonable approach here, as I think, imho, that it has better queue
> management and spooling "smarts". YMMY, of course.
Depends a lot on who actually is interacting with the physical
printer and what they're trying to do with it. If it's your normal mainframe ops
people managing the queue, then I'd tend to agree. If it's someone else, the WWW
interface for CUPS beats RSCS's interface hands down for "general user"
friendly queue management (check out http://your.linux.instance:631 on your
nearest CUPS install to see what I mean).
CUPS' IPP support also allows much more sophisticated utilization
of printer features, and mapping of LPR options and queue names to IPP option
strings (check out use of CUPS instances as preconfigured option packages for
LPR printers). The ability to trivially specify output as 2 up, duplexed,
stapled on the long side, and first and last page drawn from different paper
bins w/o coding a zillion escape sequences in RSCS -- that's easy in CUPS,
and you just map a CUPS instance name to a RSCS printer name to get it for
free.
CUPS is good stuff. And it's got a real Messages and Codes manual.
-- db
