Our experience with IPP-based, PPD-less CUPS printing

August 14, 2021

A while back I read Fedora contemplates the driverless printing future (via). This is about doing without both CUPS printer drivers (which are somewhat uncommon) and CUPS PPDs (PostScript printer definitions), and instead having CUPS interact directly with the printers through IPP. PPDs traditionally matter in CUPS printing because they tell you what the printer can do (for example, duplexing) and how to instruct the printer to actually do it. An incomplete or inaccurate PPD would either omit printer features or the features wouldn't work.

(You can read technical details about CUPS PPDs in their documentation on their extensions to the basic Adobe standard.)

One of the significant services we (normally) operate is printing, which is implemented with a CUPS server. Traditionally, finding, testing, and managing PPDs for our various Postscript based printers was an ongoing hassle, often involving a certain amount of guesswork, approximation of printer models, and substitution of one with another until everything seemed to work. When we updated our CUPS server to Ubuntu 18.04 (from 16.04) this past February, we opted to try making it a driverless server (following the Debian directions), in the hopes that this would make everything just work, especially for future printer models.

Perhaps unfortunately, our experiences with this have been a bit mixed. Some but not all of our printers worked with dynamic, driver-less IPP printing. Now, to be fair, part of this is that we have any number of very old printers, some of which were discontinued in 2005. They stay on because they work reliably and are inexpensive to operate, which is a big consideration around here (they're also monochrome only, which is a feature for departmental general use printers). Our more modern printers seem to be fine, with one exception where driverless printing had odd duplexing issues that went away when we reverted to PPD printing.

Since the past while has not exactly had high print volumes, it's possible that more issues will start emerging when more people come back to the office and begin printing more material. But we haven't had any problem reports so far.

PS: One small irritation of driverless printing in our environment is that the printer has to be up and responding to you when you add it to the CUPS server (because CUPS immediately tries to query it for its printing information). Our printers are distributed all over three buildings and people don't always leave them turned on, which apparently created some issues when my co-worker was setting up the new print server. Fortunately you can apparently repeatedly re-add printers until it works.

Written on 14 August 2021.
« Some of my views on using YAML for human-written configuration files
Chrome's automatic updates and the power it holds on the web »

Page tools: View Source, Add Comment.
Login: Password:
Atom Syndication: Recent Comments.

Last modified: Sat Aug 14 09:09:05 2021
This dinky wiki is brought to you by the Insane Hackers Guild, Python sub-branch.