Since range and array are more or less interchangeable computationally (but
not memory-wise), I suspect the main issue is how a range object should be
reported in the REPL. Julia REPL already tries to be smart about reporting
an array, either listing out all the elements or showing the beginning and
end with ellipsis in between. Perhaps it would be helpful to do something
like this:
julia> linspace(0,10,1000)
1001-element Range{Float64,1}:
0.0
0.1
...
10.0
Or some such; it's probably wrong to express Range as a container like
Array. But idea is that in the REPL, when someone queries a range or array,
they probably want to see what's in it, because they want to verify whether
the generator does what they intended. So it would help to inform the user
that the object is indeed a range, the type it contains, and a summary of
the actual values therein.
On Wednesday, September 30, 2015 at 2:51:46 PM UTC-4, Christoph Ortner
wrote:
>
> I, on the other hand, don't understand the obsession with abstraction.
>
> Also, by your argument the following should all be special types rather
> than box-standard arrays.
>
> julia> typeof(ones(10))
> Array{Float64,1}
>
> julia> typeof(zeros(10))
> Array{Float64,1}
>
> julia> typeof(eye(10))
> Array{Float64,2}
>
>
>
>
>
