Explain why we need to duplicate / shift parameters

[fingolfin]

E.g. in linear_combination

[fingolfin]

Maybe also an example would be illustrative. E.g.

julia> T = generic_character_table("SL3.n1");

julia> linear_combination([2], [T[4]])
Generic character of SL3.n1
  with parameters
    nl1 ∈ {1,…, q - 1} except nl1 ∈ (q - 1)ℤ
  of degree 2*q^2 + 2*q + 2
  with values
    2*q^2 + 2*q + 2
    2*q + 2
    2
    (2)*exp(2π𝑖((-2*a*nl1)//(q - 1))) + (2*q + 2)*exp(2π𝑖((a*nl1)//(q - 1)))
    (2)*exp(2π𝑖((-2*a*nl1)//(q - 1))) + (2)*exp(2π𝑖((a*nl1)//(q - 1)))
    (2)*exp(2π𝑖((-a*nl1 - b*nl1)//(q - 1))) + (2)*exp(2π𝑖((a*nl1)//(q - 1))) + (2)*exp(2π𝑖((b*nl1)//(q - 1)))
    (2)*exp(2π𝑖((a*nl1)//(q - 1)))
    0

julia> linear_combination([1,1], [T[4],T[4]])
Generic character of SL3.n1
  with parameters
    nl1 ∈ {1,…, q - 1}, nl2 ∈ {1,…, q - 1} except nl1 ∈ (q - 1)ℤ, nl2 ∈ (q - 1)ℤ
  of degree 2*q^2 + 2*q + 2
  with values
    2*q^2 + 2*q + 2
    2*q + 2
    2
    exp(2π𝑖((-2*a*nl1)//(q - 1))) + (q + 1)*exp(2π𝑖((a*nl2)//(q - 1))) + (q + 1)*exp(2π𝑖((a*nl1)//(q - 1))) + exp(2π𝑖
    ((-2*a*nl2)//(q - 1)))
    exp(2π𝑖((-2*a*nl1)//(q - 1))) + exp(2π𝑖((a*nl2)//(q - 1))) + exp(2π𝑖((a*nl1)//(q - 1))) + exp(2π𝑖((-2*a*nl2)//(q
    - 1)))
    exp(2π𝑖((b*nl2)//(q - 1))) + exp(2π𝑖((a*nl2)//(q - 1))) + exp(2π𝑖((-a*nl1 - b*nl1)//(q - 1))) + exp(2π𝑖((a*nl1)//
    (q - 1))) + exp(2π𝑖((b*nl1)//(q - 1))) + exp(2π𝑖((-a*nl2 - b*nl2)//(q - 1)))
    exp(2π𝑖((a*nl2)//(q - 1))) + exp(2π𝑖((a*nl1)//(q - 1)))
    0