Re: [Bug-gsl] Documentation bug(?)

[Brian Gough]

At Mon, 20 Aug 2007 13:20:34 -0500 (CDT),
Justin Lenzo wrote:
> Forgive me if I am reading things wrong, but the manual section 8.3.10
> (Vector properties) seems incorrect.  It reads:
> 
>  To test for a non-negative vector, use the expression !gsl_vector_isneg(v).
> 
> But gsl_vector_isneg() seems to return false if *any* entry is
> non-negative.  The vector being non-negative would require *all* entries
> be non-negative.  It seems one would have to test gsl_vector_min(v) >= 0
> to test for a non-negative vector.
> 
> The same sort of issue appears in the analogous matrix page.  Also on the
> matrix page, I think you might want to note on this page that one needs to
> disable the default error handler in order to use the Cholesky routine as
> a *test* of positive definiteness (the routine calls GSL_ERROR() if the
> matrix is not positive definite).

Thanks for the bug report, I'll correct the manual.

-- 
Brian Gough
(GSL Maintainer)

Network Theory Ltd,
Publishing the GSL Manual - http://www.network-theory.co.uk/gsl/manual/