XServer & Magnification: Outcome of Discussions
- together DAMAGE, XFIXES, and COMPOSITE are solving many of our
magnification issues
- notification
- interaction
- performance
- Current graphics hardware has plenty of power, problem is getting
to it using X
- X DISPLAY/SCREEN model of desktop is getting obsolete; 'session'
model is better
- XInput has problems for HIDs which may need short-term fixes
- XEvIE, XKB,(subset), DAMAGE, and XFIXES should be
standardized.
- we may be able to exceed, rather than meet, current magnification
practice
- COMPOSITE is not quite mature enough yet, needs more exposure
Actions
- Investigate pixel shaders for text smoothing, etc.
- Decide which versions of above extensions to promote in our first
FSG standard(s)
- Add API to XFIXES to separate "set cursor" from "show
cursor"
- Extend XInput to allow disambiguation of HIDs
- specify and prototype compositing-manager hints (in order to
implement some mag features)
- do proof-of-concepts of various XEvIE/mouse-coordinate approaches
Issues (also actions - to investigate and discuss further)
- Need solution for 'spurious cursor' problem before taking
advantage of COMPOSITE fully
- COMPOSITE may still be too slow for some (non-mag) users, slowing
adoption
- Input devices (HIDs) are still a problem
- need strategy for cascading/multiple or pluggable Compositing
Managers
- Should CMs be able to redirect output (to a different SCREEN, for
instance?)
- current client-side gnome-mag APIs are too GNOME-centric
- should we try to share existing 'gnome-mag' implementation source
code ?
- X security is getting tightened, which may require replacement of
some of our current mouse and keyboard manipulation techniques
- XEvIE is being pulled in several, possibly conflicting directions
- mouse coordinate transformation problem is not yet solved,
another COMPOSITE obstacle