RandR version 1.2 revisited, subpixels, point-spread function

[Keith Packard]

On Wed, 2006-09-13 at 18:38 -0400, Adam Jackson wrote:

> Of course, one obvious choice of property would simply be the complete EDID 
> block.  Right now we post a root window property whenever someone calls 
> xf86SetDDCproperties, which means the last monitor touched wins.  This also 
> has the pleasant property of detaching monitors from root windows.

Seems like a fine plan; yet another general-purpose data transfer
mechanism in the X server ftw.

> On the topic of subpixels, the DI-EXT extension to EDID has three full bytes 
> for describing subpixel layout, configuration, and shape.  The raster idea is 
> flexible but I can imagine it becoming a parsing nightmare - how do you 
> represent diagonal subpixels?  Except in very limited situations the driver 
> is unlikely to be able to tell you any more about subpixel geometry than it 
> gets from EDID anyway.

Right, pictures I've seen show us moving away from constant sub-pixel
geometries to variable (tiled in some shape) sub-pixel geometries, like
the olpc display and the hextile stuff. While it's possible to describe
those to applications, we'd also have to describe where on the screen
their pixels are landing, and they'd have to redraw when windows were
moved.

But, I'd certainly like to allow people to experiment with this kind of
stuff; general properties seem flexible enough to allow that at least.

-- 
keith.packard at intel.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://lists.x.org/archives/xorg/attachments/20060913/c3e1e988/attachment.pgp>