[Insight-developers] Re: Vector image pixel type ?

[Kris Thielemans]

Hi Karthik
 
it's probably too late by now, but I do have a comment on this thread. Feel
free to ignore it.
 
>  Image< P, 3 >
>
> It would make sense for P to have a bunch of operations defined, so 
> filters can work just as usual with it.
> 
> For instance
> P1 + P2, P1*P2, P1*scalar, etc.....
>
> I would hesitate to add these operations to itk::Array, since its too 
> general. People could for instance be using an Array of images or an 
> array of transforms. You can't define any operations here. It the same 
> reason why there is both itk::Vector  and itk::FixedArray.

In another library, I was faced with exactly the same design decision. Do I
add operator+ etc to the Array class or not? 
 
At first go, I didn't, essentially because I didn't want to compilation to
break when using a non-numeric type for P, which I believe is what you are
saying as well. Then I remembered that as all this is with templates, and
templated classes have the very useful property that their members are NOT
compiled, unless you call them (or you instantiate the whole class
explicitly obviously).
So, you can add operator+ to Array, and it won't break anything, even if you
use std::vector for P.
 
So, I reversed my first design. I got rid of my 2nd class (let's call it
NumericArray) and moved all its numerical operators to the Array class.
Result: 1 class less to worry about, less lines of code, and exactly the
same functionality. I still think that's the best thing to do.
 
PS: if you reply, please CC to my email address. I'm travelling and have
disabled the list.
 
All the best
 
Kris Thielemans
Hammersmith Imanet, part of GE Healthcare London, UK