Hacker News Reader

Comment by cosmotic | Hacker News Reader

Printing on windows and macos has been pretty stagnant for years. Windows printing is still a pretty awful experience; the print queue gets jammed and windows is really slow to unclog. The worst part is finding and updating drivers. MacOS printing does a lot better with drivers and clogged queues but it too could still use some work.

It's too bad there doesn't seem to be much industry support for a common, open driver system. PS was a good try but seems Adobe licenses killed that one. Similar with PCL and HP. There was a glimmer of hope with XPS years ago but that quickly fizzled.

View on HN

> the print queue gets jammed

Then you cannot remove the "to be printed" documents and have to kill the spool service...thats a problem since Win2K..shame on you Microsoft PUUUUUUU!

Recently at my father in laws Win10 machine. Had to get there to fix it so he can print again #2020

How is PUUUUU pronounced?

Now don't quote me on this but I think you pronounce it:

Ffs Microsoft get your shit together, windows may be dead to you but there a over a million running systems that you bought your way into. So sit up straight in your chair and get through that decade long backlog...

>How is PUUUUU pronounced?

Like Booo but more firmly ;)

Ah, spoolsv.exe my old friend.....

By a common, open driver system, do you perhaps mean "driverless"? My new printer works perfectly on Linux but took quite a bit of fiddling to get to work on Windows. Scanning from iOS is a mess (the official app works, but it has a hardcoded list of integrations instead of proper "share" support, so actually using the scanned results is a pain). I haven't even tried scanning from Windows.

And the very best part: on Fedora 32 or better, scanning Just Works! This is the result of a project to implement both Microsoft's and Apple's scanning protocols in SANE, and somehow the result seems substantially better than Apple or Microsoft's implementations.

PCL5 is well supported in my experience. PCL6 on more modern printers. Compared to Linux, the Windows printing experience is easy. 99.9% of the time, you go into your settings, click on devices, click on "add device" and wait for the printer to show up (usually within 30 seconds, sometimes takes 2-3 minutes). Then you click that and you can print. That is the 99.9% case.

On Linux you're way less lucky if you wanna print, god forbid the document is more complex than a text file.

? Connect printer, print? That's been my general experience, and we've also been printing through Linux at work for decades now - and with CUPS about since it arrived. The number of printers we've been through I have lost count of, but they work. We print all kind of documents, at sometimes an industrial scale. And of course it doesn't matter what the document is, pure text or not. At the home front the HP printers "just work", the only slightly tricky one was a Ricoh laser printer I was given - it doesn't use PS or PCL but instead its own system. That needed a 3rd party driver.

I bought a Lexmark Laser Printer a while ago. CUPS doesn't recognize this printer automatically, you have to add it manually. So you have to go through a dialog that is basically user hostile and I can't imagine how many people with less technical knowledge give up here.

Regardless of using PCL6 or the Lexmark driver from Lexmark, the printer continues to be unable to print PDFs directly from the browser. Other PDF viewers print normally. CUPS logs no error but prints it on a piece of paper instead.

When attempting to view the print queue of my printer, it crashes if a document is being sent but has not started pritning yet. It is fine if the document is queued but not printing.

Sometimes, CUPS doesn't print the document immediately but waits 30 minutes. Other documents sent to the queue in the mean time print immediately. Sometimes the program that attempts to print causes a hard lockup in CUPS, so I have to restart the CUPS service.

"Connect printer" isn't the easiest in CUPS if it decides that autodiscovery doesn't do it, which it often does.

Great anecdote, but I have at least a thousand stories like this in Windows.

My favorite one is how a guy in my office one day was not able to print. Well he could print, but only from applications that used the really, really old Windows printing API. So basically you could print from Notepad, and that was it.

We tried everything to get printing to work, but could not figure out the problem. This led to a wipe and reload of the OS. After I got it reloaded, I tested out printing, and it worked great. Success!

Next day, he comes in and is again not able to print. Turns out that the problem was due to the drivers from his personal HP printer at home. The printer was auto-detected, and Windows automatically downloaded the new driver for that printer. That somehow fucked up the entire printing subsystem again so no printers would work at all.

Printers are one of those corners of linux that still resemble the winmodem days: if it works right away it just works, but if it doesn't, you're probably out of luck. That's because drivers are key and manufacturers have little incentive to keep updated drivers in CUPS - particularly for cheaper models and cheaper brands. Industrial-grade models and brands are more likely to have had a driver done for some reason or another (big contract in civil service etc).

> It's too bad there doesn't seem to be much industry support for a common, open driver system.

Isn't Internet Printing Protocol pretty much that, and supported by most printers?

https://en.wikipedia.org/wiki/Internet_Printing_Protocol

That may have been the intent but my experience is that I still need to install drivers to get this to either work at all or work with all the features of the device.

My main issues when I owned inkjets is the printhead alignment is programmed in the drivers and when your printer gets old the printhead dont align anymore and you get a printer that still prints but pretty useless.

Even AirPrint seems to fail with pretty distressing regularity for me.

Isn't AirPrint just CUPS over Bonjour?

AirPrint is a just a brand name applied to the Internet Printing Protocol, although it uses Bonjour for discovery.

Kind of like Microsoft's brand new "Universal Print" is a brand name applied to the Internet Printing Protocol, although it uses Active Directory for discovery.

Google, recently announced they were killing CloudPrint to support IPP instead.

On the bright side, I can print reliably from Linux now, which is the first time I've been able to do that since 2005. Thanks Greg Kroah-Hartman!

The Windows printing system is a total shit show. There are at least 3 different printing subsystems under the hood that all work differently, and none of them work particularly well.

Hasn't Apple been a major adopter of/contributor to CUPS?

Apple owns the code to CUPS. They hired Sweet, the author, and got the code as well.

Own and own. They have licensed it under a permissive open source license https://www.cups.org/blog/2017-11-07-cups-license-change.htm... so everyone would be free to develop it further as they want.

A permissive but GPL-incompatible license unfortunately. Also, everyone was already free to develop it further as they wanted before the license change.

What would change in the current situation had they used a GPL-compatible permissive license?

It wouldn't change the development situation, but Linux distros wouldn't have lots of GPL violations and be ignoring those GPL violations.

Apache 2.0 software is all over the place in Linux distributions. And it's probably the license that most lawyers would recommend for a permissive license. The reason it's not compatible with GPLv2 is the patent termination and indemnification provisions (in the view of the FSF). The code isn't in the kernel though.

How is it GPL-incompatible when the CUPS project explicitly made exceptions to the Apache 2.0 License for GPL compatibility?

Linked above by @phonon https://github.com/OpenPrinting/cups

Reply on news.ycombinator.com