[Openicc] Proposed "ucmm" display profile configuration convention used by ArgyllCMS.

[Graeme Gill]

Hal V. Engel wrote:
> On my xinerama system Oyranos is able to get the correct EDID data for both 
> monitors without any issues.   I have confirmed this by looking directly at 
> the resulting Elektra configuration data.   So at least on some systems it is 
> possible although this might be driver specific (I am running the nvidia 
> driver right now).

I couldn't see any mechanism for retrieving the EDID on the system I tested.
The root window property only contained the EDID for the first display, as
Xinerama didn't seem to pass on properties from the second display. Maybe
this has changed in more up to date software, but I didn't locate
any documentation as to the expected behavior.

>>Unfortunately I've found it's not easy to test XRandR 1.2 at the moment.
>>The Intel driver seems to be the only one supporting it, and even though I
>>specifically bought a Motherboard with an Intel graphics chipset to be able
>>to test it, the motherboard is rather crippled: There is only one Video
>>output, and as soon as you plug another video card in, the BIOS turns the
>>Intel graphics off. The manufacturer seems to think this is a "feature",
>>aren't interested in fixing it so that you could use the on board graphics
>>with a plug in card). 
> 
> 
> I think these Intel chip sets expect you to plug in an interface card (not a 
> video card) that is designed to supply the additional video ports.  None (or 
> almost none) of the motherboards with these video chip sets actaully have 
> second video port built in. 

I'd be happy if it had this. It doesn't though, there is no connector to
add a interface card for a second video out.

> The Nouveau driver (open source driver for nvidia cards) is now XRandR 1.2 by 
> default and it appears that support for the NV4x series of GPUs (IE. 6xxx and 
> 7xxx cards) is getting farily good now.  The developers claim that in 2D it 
> outperforms the vendors blob and much of the 3D stuff is working for the NV4x 
> GPUs.  But it is at best alpha level code and is also dependant on 
> developement versions of the kernel, xorg and Gallium.  So getting this setup 
> working would be a significant task and it would be difficult to keep it up 
> to date.

I'm finding the learning curve rather burdensome for non-distro drivers. It's easy
to waste a week or more trying to get such stuff going, and I really don't want to
spend time learning all about GIT, building X servers etc. etc., since this is not
my focus.
(I wasted over a week playing with Gentoo and Debian, without getting either going
  on my rather vanilla system. Life is too short, and there is too much else to do.)

> I beleave the the RandonHD driver (the new open source driver) is also XRandR 
> 1.2 at this time.  This should be easier to get working than the Nouveau 
> driver because it is closer to being beta level code at this point.   Again 
> the xorg list should have developers working on this that can clarify which 
> version(s) of this driver support XRandR 1.2.

One can't count on getting a response though in my experience (e.g. my
report about problems getting the VideoLUTs contents via XRandR 1.2).

Graeme Gill.