On 7/26/05, Russ Cox <[EMAIL PROTECTED]> wrote: > Modern Unix printing is a catastrophe. Plan 9 printing is only a mess.
What's your take on IPP? I've done CUPS -> Plan 9 and vice versa (the latter being easier) via LPD, but I've always found that IPP in any of its flavors seems to be less painful (though MS clearly has trouble thinking out of the box on this one). My big thing is that CUPS specifically does a decent job of tackling the two problems with any printer: - What the hell can it do - What format do I send the data Its answer is to get the PostScript Printer Description via HTTP, then (ideally) either handoff PostScript or use GhostScript to translate the PostScript into the proper page description language and then hand it off. It will do raw print queues, too, which can be handy when you have Windows clients or need to manage PCL-only printers. The big upside is that for systems that don't share a namespace, you can pretty easily share printers and printer configurations and still have client- or application-specific preferences that take advantage of printer-specific features, without the sort-and-staple queue, double-sided-color queue, etc. Server-side can be a mess (and occasionally a catastrophe), but client-side is extremely clean. > but maybe the _ps suffix is common elsewhere. If memory serves, Brother, Canon and the old HPs tend to have local LPD queues specific to the page description language, so you'll see _ps and _pcl suffixes. Usually there's a queue that's just "lp" that will auto-detect, but some people like to use the other named queues to filter jobs that aren't in the correct PDL, especially where buggy applications/jobs can cause the printer to behave poorly. -Jack
