Proposed F19 Feature: CUPS 1.6

[Chris Murphy]

On Jan 30, 2013, at 7:13 PM, Benjamin De Kosnik <bkoz at redhat.com> wrote:

> 
>> Quality evaluation needs test targets/documents, and eyeballs. 
> 
> I think you mean "trained eyeballs" above.

I guess in some cases that's useful. But surprisingly a lot is tested without anyone looking. If they had, they'd have had a "this isn't right" moment.

> All these fancy color-calibration mechanisms in
> gnome/cups/ghostscript are useless for me unless one-color printing is
> first correct (ie page size/resolution).

Well that's a small problem as inkjet printers are extremely non-linear, and have n-color channels. So most manufacturers have moved to treating them only as RGB, with no access to single color. For now the professional printers, there's access to n-color channels, but can be tedious to setup.

> 
> LOL. Agreed. 20 years. Linux clearly has the capability to be much
> better.

The primary cause for the problem on OS X, is ideologically impossible on Linux. It's practically impossible on Windows. So on two normally inhibiting fronts to making bad choices, Apple has overcome the odds. It wasn't this way circa OS X 10.3 and earlier.

A problem Linux has had is constraining the behavior of inkjet printers to something reasonable, while also not exposing dangerous and mutually exclusive options in the print (driver) dialog GUI. I haven't looked at this in Fedora yet.


Chris Murphy