Change [] to () for function evaluation about approxfun.jl HOT 6 CLOSED

Should f[k] be interpreted as a vec(f)[k] when k is an integer for the time being?

from approxfun.jl.

I think that is cool. Out of curiosity, how expensive is vec(f) ?

from approxfun.jl.

It should be as expensive as making a copy of the coefficients, so not very. (It just deinterlaces the coefficients.)

On 2 Dec 2014, at 6:15 pm, Gustavo Goretkin [email protected] wrote:

I think that is cool. Out of curiosity, how expensive is vec(f) ?

—
Reply to this email directly or view it on GitHub #52 (comment).

from approxfun.jl.

Is there a reason not to store the coefficients in separate arrays in a vector-valued fun?

from approxfun.jl.

Right now the Fun type requires that the coefficients are a Vector{Float64} or Vector{Complex{Float64}}. This is consistent with the \ operator, which always works on infinite-length vectors. Doing special storage for different Fun types would complicate things

On 2 Dec 2014, at 6:26 pm, Gustavo Goretkin [email protected] wrote:

Is there a reason not to store the coefficients in separate arrays in a vector-valued fun?

—
Reply to this email directly or view it on GitHub #52 (comment).

from approxfun.jl.

Ah, that makes sense. An operator on a vector-valued fun also has its rows interlaced in the same way then, right?

from approxfun.jl.