Re: [boost] [Review][PQS] Review ends

From: me22 (me22.ca_at_[hidden])
Date: 2006-06-10 14:07:12

Just a few observations...

On 6/10/06, Andy Little <andy_at_[hidden]> wrote:
> IMO it makes sense to put 3D entities in a 3D space and 2D entities in a 2D
> space, but its not critical to me either way
>
Can the vector not be templated on length?  I know a simple
(dimensionless) vector can be templated on length (
http://gpwiki.org/index.php/C_plus_plus:Tutorials:TemplateVector ),
but perhaps PQS needs something that prevents this...

> >    I believe that in matrices there is no need to support both row-column and
> > column-row notation explicitly, but rather to keep internal compatibility
> > thorough whole library.  Based on that I'd strongly suggest to rename
> > rc_matrix
> > into matrix, and never implement cr_matrix. There is an interesting reasoning
> > about
> > this problem inside panda3d software manual (given time I'll find it later).
>
> Its certainly less work sticking to one type of matrix so I dont object to that.
> There may be performance resaons to choose on or other type?
>
If the internal storage for the matrix is usable as an array,
interfacing with different C apis would be much easier if the ordering
could be chosen (or at least relied upon).  As I recall, OpenGL uses
column-major and DirectX uses row-major, so there could be big savings
in not needing to copy the values when passing the data to those APIs.