[BUG] Misleading presence of `normalize` keyword in `orthogonalize!`

[VinceNeede]

Description of bug

Function orthogonalize! still accepts normalize keyword even though the normalization is not performed and no warning is printed. I saw the TODO, but while the maxdim is still used so the truncation is still performed, normalize is not used anymore.

Minimal code demonstrating the bug or unexpected behavior

Minimal runnable code

julia> using ITensors, ITensorMPS

julia> s = siteinds("S=1/2",2)
2-element Vector{Index{Int64}}:
 (dim=2|id=429|"S=1/2,Site,n=1")
 (dim=2|id=23|"S=1/2,Site,n=2")

julia> ψ = MPS(s, ["0", "1"]) + MPS(s, ["1", "0"]);

julia> orthogonalize!(ψ, 1; normalize=true)
MPS
[1] ((dim=2|id=429|"S=1/2,Site,n=1"), (dim=2|id=332|"Link,l=1"))
[2] ((dim=2|id=23|"S=1/2,Site,n=2"), (dim=2|id=332|"Link,l=1"))


julia> norm(ψ) ≈ 1.
false

Expected output or behavior

Either obtaining the state normalized but with a depwarn or an unexpected keyword error.

[mtfishman]

I honestly can't remember what the plan with that was, seems like we should just remove it as an input and have it error if a user passes it.

[VinceNeede]

I agree with you, the main problem I think is the fact that the keyword is used in the tests, so they should be changed too

[VinceNeede]

@mtfishman If it is okay for you, I can attempt to modify it, but I wanted to ask you what did you wanted to do about the maxdim keyword since in the TODO it was also said that it should be removed. So let me know if you want to leave it like this or either add an ITensors.warn_once or delete the keyword

[mtfishman]

Let's keep maxdim for now to avoid making a breaking change.