Re: Image transformations

[Stefan Monnier <monnier@iro.umontreal.ca>]

> I don't think I understand what "display position" means here.  Did
> you mean coordinates of a pixel?

Yes, I meant the coordinate in the display as opposed to coordinates in
the original image.

> What actually happens AFAIU is that in some cases the matrix describes
> how to convert the original image to the transformed image, while in
> others it describes how to transform the image space into the
> transformed space.

That's what I meant to say, yes.

>> Also, one of the two matrices can't be computed if the transformation
>> turns the 2D image into a line (or a point).
> We don't support such transformations.

There's no option to "magnify" by a factor 0?

>> Compared to the cost of handling the image, inverting a 3x3 matrix
>> is probably negligible, tho.
> Maybe so, but it still sounds silly to me.

Part of the reason for my question above is to figure out if the
implementation of the API is likely to have such matrix inversion
code internally (since it seems quite possible that sometimes one is
preferable and sometimes the other).


        Stefan