Difference in pmap and map

[Rajath Shashidhara]

Hello,

I'm trying to understand why pmap and map do not follow the same convention. 

[1] map retains the type in the result but pmap does not.

[2] pmap flattens the n-dimensional arrays before applying the function.

For e.g.:

julia> x = rand(2,2)

2x2 Array{Float64,2}:

 0.79318   0.368926 

 0.878642  0.0445509

julia> function f(x):

           2*x

       end

f (generic function with 1 method)

julia> map(f,x)

2x2 Array{Float64,2}:

 1.58636  0.737852 

 1.75728  0.0891018

julia> pmap(f,x)

4-element Array{Any,1}:

 1.58636  

 1.75728  

 0.737852 

 0.0891018

Is there any rationale behind it? It would be nice if both behaved in the exact same way.

Thank you,

Rajath Shashidhara