Mindboards

I know how you see it, but if you do so and you consider it to be "correct", then you must be able to calculate with the values of your model and your model has to give mathematical correct results.

If you look at a square like
::
as it should have a length of 2 and a height of 2 (in the language of your model, 2x2):
how long is the diagonal in your model?

how long is the diagonal of a rectangle like
::::
(2x4 in your model)?

Pixels are made to be visual points. When together with other pixels, the eyes perception is the perception of a filled polygon. Then, it makes sense to see a pixel as a 1x1 square, for me. Otherwise, I wouldn't even see a single alone pixel, but I see it.

Anyway, if it is a matter of FW compatibility/simplicity, I don't really mind to have it like that. But I know that if I only use the function again next year, I will do the same mistake.

yes, but if you got 10 pixels side by side ( for you: a 10x1 rectangle = 10 square pixels area extension) : what makes the difference to a 10 pixels long LINE (no area extension) ?

As I said before, one may describe and define for oneself what one likes (also mathematicians don't do it other way), but it has to be mathematical coherent and consistent.

There is nothing different between a 10x1 rectangle, and a 10 pixel line when a pixel is the same as a 1x1 rectangle, not just a point.

We all understand your point of view, so stop trying to convince us differently. No matter how un-scientific you think our views are, just leave them to our-selves. If for no other reason, it's already like you want it to be, so let us be the ones that need to remember to convert it into the dimensions we see.