[cairo] Re: SVG backend patch - update

Jason Dorje Short jdorje at users.sf.net
Sat Sep 3 17:20:52 PDT 2005

James Cloos wrote:
>>>>>>"Jason" == Jason Dorje Short <jdorje at users.sf.net> writes:
> Jason> 90 DPI at arm's length is still 90 DPI.  dots-per-inch is
> Jason> distance-independent unlike dots-per-arcsecond (for instance).
> The spec recomends that the pixel size be based on arcseconds,
> specifically the number of arcseconds subtended by 1/90 inch
> at a radius of  approx 28 inches.
> Jason> (although one guy on #svg said the spec was misworded and the
> Jason> correct value would become 96 DPI in a future spec).  This is
> Jason> also what inkscape uses (not configurable).
> Isn’t 96 what cairo itself uses?  Sounds, then, like the best bet.

libsvg used to use 100.  This was changed to 90 after some guys on #svg 
pointed me to the spec.  Admittedly my main impetus for this change was 
that using 100 DPI made many SVG files created with sodipodi/inkscape be 
rendered incorrectly...but this is a bug in inkscape/sodipodi which 
mixes units (pixels/inches) in its files.

If SVG files were correct, it wouldn't matter what DPI we used.  So for 
the cairo SVG backend, as long as the output files are correct, I guess 
it doesn't matter what DPI is used...


More information about the cairo mailing list