[Openicc] _ICC_PROFILE in reality

Kai-Uwe Behrmann ku.b at gmx.de
Thu May 26 03:16:26 PDT 2011

_ICC_PROFILE in X is a defacto standard adopted by very many 

Am 26.05.11, 10:30 +0100 schrieb Richard Hughes:
> At the moment gnome-color-manager does the following:
> 1. Sets the output atom _ICC_PROFILE on each xrandr output
> 2. Sets the screen atom _ICC_PROFILE only for the primary output
> GCM doesn't do the whole _ICC_PROFILE_1 thing (where _1) is the
> xinerama ID for four principal reasons:

Then multi monitor aware clients might then not work. GCM is then
not compliant to any _ICC_PROFILE in X spec. The XRandR stuff was added 
in version 0.4, which is the actual one.

> * Nothing exists that reads anything other than _ICC_PROFILE

I have seen Inkscape swapping colours, while moving over the monitor 

> * We're living in a post xinerama world, and applications that are
> using xrandr have no way of mapping a xrandr ID to a xinerama ID.

You might refere to the server side xinerama. Then thats true and at the 
same time non relevant to clients and normal users. The client side 
xinerama API will stay for a long time.

> So, this is the situation I'm in: I've been asked to merge the
> majority of the session components of gnome-color-manager into
> gnome-settings-daemon. I don't want to merge things that are not
> useful, and so far I see the following things as not useful:
> * The per-output _ICC_PROFILE atoms -- they make logical sense, but
> *no* applications use them
> * The _ICC_PROFILE_1 thing -- it doesn't exist in GCM, and no apps
> seem to use it.
> Does this seem sensible?

Not all clients can easily be identified to need a certain feature.

Basically it boils down to,
if we want to get Linux better accepted as a relyable platform, we should 
keep existing specs working and provide compatibility as long as possible.
Thats why I think it would be a good thing for Gnome and Linux to fully 
support the _ICC_PROFILE in X spec.

Kai-Uwe Behrmann
developing for colour management 
www.behrmann.name + www.oyranos.org

More information about the openicc mailing list