[Openicc] Is CUPS the right place for printer color management...

[Chris Murphy]

On Feb 7, 2011, at 6:48 PM, Robert Krawitz wrote:

> On Mon, 7 Feb 2011 16:04:09 -0700, Chris Murphy wrote:
>> So considering this, I would not reinvent the wheel. I don't see a
>> problem with *cupsICCProfile. It enables exactly this future
>> potential.
> 
> Except that it's an absolute pathname that's system-specific, and has
> a very limited number of selectors that can be used (why, oh why is a
> limit of three selectors baked into that kind of spec????).

Look at who wrote that spec and for what market. Those are the only things that the manufacture producing both the driver and the ICC profiles use to differentiate. It was custom tailored for what already existed at the time Apple bought the source code and incorporated it into OS X.

Sorry but the Gutenprint drivers are like an LSD trip. And I've never done LSD. But I see all of the options and I feel like I'm hallucinating. That kind of control is very cool, but I will bet that even in the Linux world maybe 1% of the user base takes advantage of this and the rest use presets. I am fairly masochistic in trying to understand things at a lot of detail, but the Gutenprint drivers are so extreme on the amount of detail that I can't even imagine what kind of parameters you'd need to be able to pass with *cupsICCProfile to attach to various profiles. It would be really extraordinary if you're going to account for all of those settings.

But can't you fork CUPS and add the additional selectors you want? I do agree that absolute pathnames are B.S. They should be URIs.


Chris Murphy