document the difference between float32 and Float32, etc.

[mgautam98]

Solves #139
Added documentation to advertise the usage of lower-cased forms of Float32, Float64 etc

[johnnychen94]

Thanks for doing this! I opened that issue as a memo to myself but apparently you're handling it rather eagerly.

It's really nice to see these improvements in the documentation.

[timholy]

This is a great idea, thanks for tackling this.

[timholy]

"Works as expected" depends entirely on what your expectations are. Some people will expect it to return a single floating-point number.

How about copy/pasting the examples with a single colorant object that I gave in my explanation to you? This mixes together broadcasting with colorspace conversion, and we don't have to mix the concepts.

[johnnychen94]

I'd like to also add the following table into the document as a cheat sheet:

The following table tells you how you can convert from a number/colorant to another number/colorant

from \ to	N0f8	Float32	Gray{N0f8}	Gray{Float32}	RGB{N0f8}	RGB{Float32}
N0f8	-	float32(c)	Gray(c)	Gray{Float32}(c)	RGB(c)	RGB{Float32}(c)
Float32	N0f8(c)	-	Gray{N0f8}(c)	Gray(c)	RGB{N0f8}(c)	RGB(c)
Gray{N0f8}	N0f8(c)	Float32(c)	-	float32(c)	RGB(c)	RGB{Float32}(c)
Gray{Float32}	N0f8(c)	Float32(c)	n0f8(c)	-	RGB{N0f8}(c)	RGB(c)
RGB{N0f8}	N0f8(Gray(c))	Float32(Gray(c))	Gray(c)	Gray{Float32}(c)	-	float32(c)
RGB{Float32}	N0f8(Gray(c))	Float32(Gray(c))	Gray{N0f8}(c)	Gray(c)	RGB{N0f8}(c)	-

• float32(RGB(c)) works equivelently to RGB{Float32}(c), and their performances are the same, too.
• the lowercase function also works if c is a type, for example, float32(Gray{N0f8}) gives you Gray{Float32}
• To handle image data, you just need to broadcast it, for example, float32.(img)

Or maybe a simplified version by removing the explicit construction ones:

from \ to	N0f8	Float32	Gray{N0f8}	Gray{Float32}	RGB{N0f8}	RGB{Float32}
N0f8	-	float32(c)	Gray(c)	-	RGB(c)	-
Float32	n0f8(c)	-	-	Gray(c)	-	RGB(c)
Gray{N0f8}	-	-	-	float32(c)	RGB(c)	-
Gray{Float32}	-	-	n0f8(c)	-	-	RGB(c)
RGB{N0f8}	N0f8(Gray(c))	Float32(Gray(c))	Gray(c)	-	-	float32(c)
RGB{Float32}	N0f8(Gray(c))	Float32(Gray(c))	-	Gray(c)	-	-

I personally like the simplified version because it shows all the convenient way of converting things in JuliaImages.

@timholy will the usage of RGB(1.0) being deprecated? I believe it will still be available?

[mgautam98]

@johnnychen94 this cheat-sheet is really helpful.