RandR version 1.2 revisited

[Graeme Gill]

Keith Packard wrote:

 > Yeah, I think we've got enough functionality for version 1.2. If we are
 > going to have this kind of communication visible to applications, it
 > might be nice to wrap it up in a pretty package instead of exposing them
 > to raw monitor communications.

Judging by the type of approach that DDCcontrol has adopted,
I suspect that wrapping it up in a prettier package than the
raw DDC/CI bytes might be difficult. It seems that every
model of monitor has its own peculiarities, and having to constantly
update things to allow for this, might make the problem of parsing
DDC/CI traffic intractable for X11 driver implementors.

A more pragmatic approach might be to separate the means of
identifying the device and transporting the DDC/CI information,
from it's detailed interpretation. If the latter is left up
to the application, then there is room for a supplemental
user side library to deal with keeping a database of different
monitor peculiarities, and translating raw bytes to/from
higher level functionality. If this part becomes much
better understood at some stage (perhaps with some standardized
descriptor available from monitor manufacturers, or other
projects such as DDCcontrol), then perhaps it could then be
absorbed into the X11 side.

cheers,

Graeme Gill.