On Tue, May 5, 2009 at 4:17 AM, Casey Duncan <ca...@pandora.com> wrote:
> Just to clarify my position, performance is a secondary concern, primarily
> I'm concerned about intuitiveness and convenience. I suspect an immutable
> vector type will cause a regular amount of mail traffic asking how to mutate
> them and why that is not possible. I could be wrong, but I think many folks
> will be surprised by an immutable vector class.
>
> I'm +0 on immutable vectors, they're certainly much better than nothing,
> but I personally would prefer mutable.
>
> To make immutable types more performant, you should look at the
> implementation of tuples in Python. tuple instances are pooled so they can
> be recycled, saving a lot of memory management overhead. I did a similar
> thing for vectors in Lepton with good results.
>
> -Casey
>
+1 for memory pools. Memory is the main overhead for performance these
days. Anything to avoid a copy or a malloc/mmap, and to group areas of
memory together for reading.
Mutable vectors fail the use cases:
- use Vector as a proxy for some values in a larger array.
eg, in numpy I can do reference_to_big_array = bigarray[1:3]
- consistency with Rect, list, numpy/Numeric arrays, python arrays.
Another use case is consistency with most other Vector types in CS. I think
most Vectors are mutable.
- allow people to use vectors like how they are used in most graphics
literature.
cu,
