[Too much math before my first coffee in the morning.

Anyway, the harder I try to write a rebuttal, the harder it gets. Now, assuming a .1mm gap (1/32 inches) and 500nm visible light, the article translates to something like ~A 200 wavelengths single slit can be approximated as two semi-infinite screens~ that makes a lot of sense.

But I may be missing something.]

---

Back to your questions:

> - If the interference pattern was explained by diffraction by a semi-infinite plane, why don't I see it when using only one finger? I only see a blurry shadow. The second finger is needed to make the pattern appear.

I got the best results with a led lamp on the wall of my kitchen. It's a 5x5 inches square, with a white translucent plastic and a metallic frame that hides a led strip and can be bought in any electricity store near my home. I was standing like 10 foots away, in a slightly dark area (or use the other hand to cover the eyes and fingers).

I almost put my finger like 3 inches away from my eyes, I close the fingers until I see the pattern. Then I open them slightly and I see a phantom line around my finger that does not disappear when they are far away from each other. If I look very carefully, I see a second phantom line.

- All formulas that are used compute the light intensity projected on a screen. In the actual experiment, we're looking at the slit through a lens (our eye or a camera). That's not the same thing.

A standard trick is to use a lens and a screen at the focus distance instead of a screen at infinity. This is similar to the eye. The lens in the eye will have a different adjustment to make a clear image of the eye, so it will not be exactly equivalent. But it's close enough. I'd not worry too much about this part.

> - The fact that this is white light interference is handwaved away. To model it correctly, you'd need to compute what happens at each wavelength, then integrate the resulting spectrum multiplied by CIE's x, y z functions at each point, and finally do a bunch of math to bring that in the sRGB color space if you want to display the model's result on a screen.

To get a nice rainbow interference pattern, you probably need a almost puntual source of light. A diffuse source will make a blurred rainbow that is impossible to notice. But I need a diffuse source like the lamp in my kitchen to notice the dark lines because they are too weak.

White leds use phosphorus to get the full spectrum, I'd try with a lamp with leds of color that are almost monochromatic. I'd try with red and blue to maximize the distance of the interference lines, but I'm not sure if the blue leds use phosphorus too. Perhaps red and green is better for this. (Perhaps a computer screen with Painbrush filled with #FF00FF is a good alternative.)

-I do see it on the edge of a single finger.

-I agree, also I can only observe the effect when focusing on the gap

-sure, but weirdly the effect has to be wavelength dependent, but there are no color fringes.

I think this is something else, but haven't figured it out yet.