Re: Backend and non-backend (was Re: Stencil bounding box)

[Johannes Schindelin]

Hi,

On Mon, 1 May 2006, David Feuer wrote:

> On 5/1/06, Han-Wen Nienhuys <address@hidden> wrote:
> 
> > it should take geometry and appearance into account, ie. where the
> > symbols end up on the page, and how large they are. What I would really
> > want is to have a similarity measure for the image produced by two PDF
> > files, but that's a rather hard problem, and it will be easier if we
> > take advantage of the fact that we can access the backend.
> 
> Well, if you care about where things are on the page, one option would
> be just to compare PNGs.  Otherwise:  the new backend design I'm
> considering would flatten stencils.  This should make it really easy
> to pull out a list of elements, sorted in any convenient fashion.

You cannot just compare PNGs. The really important information is: where 
is the object, not: what color is that pixel. Just think about the 
consequences of a small adjustment of the position. The difference of the 
PNGs is potentially quite large, even if the difference of the location is 
quite small.

And especially for tasks like refactoring, it is invaluable to have 
automatic regression tests _before_.

Ciao,
Dscho