[Openicc] Linux CM ideology .. Linux CM proposal

Graeme Gill graeme at argyllcms.com
Mon Feb 7 16:48:02 PST 2011


Jan-Peter Homann wrote:
> I disagree, that the problems we have with the implementation of color
> management are based on the ICC model. I believe, that is possible to
> implement ICC based colormanagement in way, that fits the needs of
> users, which don´t want to deal with profile creation and gives the
> power users the possibility to enhance the workflow with individual
> created profiles.

Agreed. pointing to the ICC standard as the reason there is poor
support for color management on Linux, is a total red herring. While
there are many aspects of current color management technology
one might wish was different or better, there is nothing technical stopping
the implementation of something that will be effective and usable.

> Indeed, the main problem we have currently on Windows and Mac OSX is a
> "chaotic" implementation of ICC based color management on system level,
> in various applications and printer drivers, which don´t interact, or
> which interacts in non wanted ways.

Be that as it may, the anarchic nature of open source and Linux system
development makes it even harder to arrive at something that is pervasive.

MSWin and OS X had the benefits of dictate - someone in Microsoft and Apple
decided that a system wide color management system was going to be provided,
and that it was going to be used, and has made it happen. If they
had simply left it up to their developers, I would guess that neither of those
system would have color management yet either, since for the average developer,
even those dealing with graphics, "RGB is RGB". Nothing comparable is likely to
happen with Linux. The best scenario is that someone solves the problem
so well and so conveniently, that all those needing color management adopt
it, rather than re-inventing the wheel. In some areas we're off to
a good start (lcms, although there are still notable problems areas, such
as FireFox), in other areas not (Oyranos vs. colord vs Argyll's dispwin -I, etc.).

> Here at the OpenICC, we have the possibilty to work on better integrated
> solutions. The proposal i have posted at OpenICC wiki
> http://www.freedesktop.org/wiki/OpenIcc (point 9)

I would actually suggest a much more modest proposal to start, and
that is providing a system wide registry of devices and associated
profiles, and a means to read and write the registry.

I would propose creating a file format and a library to access
the settings. Make it (relatively) self contained, so that
it doesn't imply bringing a lot of baggage along, or any particular
environment or system dependence, so that it can gain wide acceptance.
It would provide a lowest common denominator, and provide
a foundation for applications to do their own color management,
as well as system libraries to provide color management
services.

I suspect agreeing on even something as simple as the above (with
the aspects of how to store it, how to access it, how to identify
devices, and how to identify appropriate profiles for a given
device state + desktop/distro/operating system politics), will
be a formidable enough challenge. If we could get that far, the
rest might flow on.

Graeme Gill.


More information about the openicc mailing list