[Insight-developers] itkVector, itkPoint and vnl_vector

[Bill Hoffman]

How do you do this without using virtual functions?

At 10:21 AM 1/4/01 -0500, Luis Ibanez wrote:

>Bill Hoffman wrote:
> > Are there any suggestions on how we can fix the vnl_vector_fixed and
> > maintain compatibility with the current vnl_vector and vnl_vector_fixed
> > classes?
>
>
>I think that Jim's proposal is a good idea.
>having a common vnl_vector class without data,
>and deriving a fixed and dynamic versions from 
>it, like:
>
>  vnl_vector<T>
>|
>|-----> vnl_vector_dynamic<T>
>|
>|-----> vnl_vector_fixed<T,n>
>
>
>What they have now includes a vnl_vector_ref that allows to use
>user allocated data inside the point (that don't seems to be very
>useful for the small amount of data of a 3D point)
>
>
>vnl_vector<T>
>      |
>      |
>vnl_vector_ref<T>
>      |
>      |
>vnl_vector_fixed<T,n>
>
>
>---
>
>
>BTW, vnl also have a matrix hierarchy, 
>that is similar to the one for vectors:
>
>
>vnl_matrix<T>
>      |
>      |
>vnl_matrix_fixed_ref<T,m,n>
>      |
>      |
>vnl_matrix_fixed<T,m,n>
>
>
>
>Maybe that should be adjusted too.
>
>
>Luis
>
>_______________________________________________
>Insight-developers mailing list
>Insight-developers@public.kitware.com
>http://public.kitware.com/mailman/listinfo/insight-developers