Re: why do we still have stdnormal_{cdf,inv,pdf,rnd}?

[Robert T. Short]

A comment on this old message.

The stdnormal_xxx functions are for zero mean, unit variance (normal) 
random variables.   The normxxx functions are for the general case but 
call the stdnormal_xxx functions.

Deprecating the stdnormal_xxx functions would break the normxxx functions.

Bob
--
Robert T. Short
PhaseLocked Systems

John W. Eaton wrote:
> On 19-Feb-2009, Søren Hauberg wrote:
>
> | tor, 19 02 2009 kl. 14:49 -0500, skrev John W. Eaton:
> | > Does anyone know if there is some reason that these functions have not
> | > been moved to the deprecated directory yet?  It seems that Matlab uses
> | > the names normcdf, norminv, normpdf, and normrnd, and Octave also has
> | > those now.  So unless anyone objects, I think we should move these to
> | > the deprecated directory before the 3.2 release.
> | > 
> | >   stdnormal_cdf
> | >   stdnormal_inv
> | >   stdnormal_pdf
> | >   stdnormal_rnd
> | 
> | It makes sense to do so.
>
> Oops, I started to move the stdnormal_X functions to the deprecated
> directory, but then noticed that they are much different from the
> corresponding norm_X functions.  Since I don't know why they are
> different, I'm leaving them in place.
>
> | Out of curiosity: what happens to functions in the deprecated directory?
> | I assume they are removed at some point. For how many releases do they
> | stay? Are there any functions that should be removed for 3.2?
>
> When functions are moved to the deprecated directory should have a
>
>   ## Deprected in version M.N
>
> comment added to them noting the version of the next major release.
> I think we should remove them after they have been in the deprecated
> directory for some number of major releases (2? 3?).
>
> jwe
>
>
>
>