Jim Gettys jg at
Thu Apr 9 08:10:36 PDT 2009

On Thu, 2009-04-09 at 10:58 -0400, Behdad Esfahbod wrote:
> On 04/09/2009 10:55 AM, Jim Gettys wrote:
> > On Thu, 2009-04-09 at 10:48 -0400, Behdad Esfahbod wrote:
> >> On 04/09/2009 10:37 AM, Alan Cox wrote:
> >>
> >>> In the perfect world the monitor would report the DPI reliably to the
> >>> Xserver which would adopt it and everything would just work.
> >>> Unfortunately monitors report rather varied things, the dominant OS
> >>> product appears to ignore the monitor value entirely (so the monitor
> >>> doesn't get corrected) and everyone forces an OS dependant DPI in their
> >>> desktop ignoring all the monitor info.
> >> All we want is a community call to gather HAL .fdi info for all common display
> >> devices...
> >>
> >
> > Again, this isn't sufficient; you also have to know the (typical)
> > viewing distance of the screen.
> >
> > Any such call should be made in such a way we get enough information to
> > actually be useful....  DPI itself isn't enough...  We need two pieces
> > of information, one of which is a nice simple number, the second of
> > which is at best an approximation.  But without the second, the first
> > isn't very useful.
> Right.  We discussed this all in detail a few months ago...
> Either ask for device-type (handheld, monitor, TV, projector, ...), or simply 
> ask for dimension-divided-by-typical-distance.

Since when you are doing certain few applications (actual WSIWYG layout,
or the use of the screen as a ruler, which we do on OLPC), you really do
want the actual DPI; it's just we *also* need to know the typical
viewing distance.....  

And in the past, we never had the viewing distance in X11's design: our
presumption was then always desktop monitors, viewed at the "usual"
desktop distance).  It is entirely missing from the core protocol.

We've been missing one value from day one, and the one that actually
makes DPI make sense for 98% of applications we use on a day to day
                           - Jim

Jim Gettys <jg at>

More information about the xorg mailing list