[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
## Re: [Help-gsl] gsl_vector_alloc( 0 )

**From**: |
Michael Lehn |

**Subject**: |
Re: [Help-gsl] gsl_vector_alloc( 0 ) |

**Date**: |
Tue, 10 Jan 2012 23:13:49 +0100 |

Am 10.01.2012 um 18:03 schrieb John D Lamb:
>* Is there any particular reason why gsl should abort on*
>* gsl_vector_alloc( 0 ) and gsl_matrix_alloc( 0 )?*
>* *
>* It is straightforward to construct a gsl_vector or matrix of size 0. And*
>* I think all the vector and matrix operations behave appropriately on*
>* such as vector or matrix: that is, they do not abort if the operation is*
>* meaningful, even if the result is undefined.*
>* *
>* There are occasions when it is useful to allow a vector of size 0. For*
>* example, it is easier to put a gsl_vector* in a struct and interrogate*
>* it with gsl_vector_size( s.vector ) than with s.vector == 0 ? 0 :*
>* gsl_vector_size( s.vector ).*
>* *
>* -- *
>* John D Lamb*
>* *
>* *
I agree. In LAPACK vectors of length zero or matrices with zero rows
or columns are used/allowed because it often eases the implementation
of numerical algorithms.
Michael