Hecate said:
They are physically rendered on a computer screen - where they have a
width and height expressed in pixels.
But that width and size of the image is not unique - on another screen
exactly the same pixels will produce an image of another size, so the
size is not an intrinsic parameter of the pixels. Just because a
rendered image has a width and height does not mean that the pixels
themselves have - it is the rendering process defines the width and
height, not the pixels. The same pixels can produce an image the size
of a postage stamp or half the surface of the planet - indeed, a
significant portion of the visible universe if necessary!
Lets make it simple for you:
The image has n x m pixels. It is rendered at y "pixels per inch".
The image height is therefore n/y pixels/(pixels/inch) = n/y inches.
The image width is m/y pixels/(pixels/inch) = m/y inches.
ie. both width and height are determined from the calculation of the
following units: pixels/(pixels/inch) = pixels/pixels * inches.
The dimensions of a pixel can be anything you like - oranges, elephants
or chocolate cookies, because they completely cancel out in all
calculations. Thus, invoking Occam's razor demonstrates they need be
nothing at all - completely *dimensionless*.
Now come up with a derivation of the dimensions you believe pixels to
have, or just apologise to Toby for getting it wrong.
I'm talking about images, I
don't know what you're talking about.
So why did you dispute Toby's statement that the pixel count is
dimensionless - you may be talking about images, but you are talking to
yourself! As Toby stated, the pixel count *is* dimensionless (which is
the specific statement you disputed) simply because it is the total
number of items which each have dimensionless units.