commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Remi Arntzen" <remi.arnt...@gmail.com>
Subject Re: [Commons-Math] FFT Support
Date Thu, 13 Jul 2006 20:38:17 GMT
On 7/4/06, Phil Steitz <phil.steitz@gmail.com> wrote:
> Hi Remy,
>
> The test coverage is still not where we would like to have it and we
> may want to refine/revise the API.  Please have a look and if you have
> suggestions for improvement or are willing to work on adding more test
> cases / validation with other packages, thanks in advance.
>

Well now that you mention it the current FastFourierTransformer class
is not public for me to use.  The transform method is public, but it
is also non-static and thus requires an instantiation of the class.
However the class is non-instantiable, because the constructor for the
class only has package level access.

I want to start work on a multi dimensional extension to this class,
but it would be somewhat easier if the constructor was made public.

Along the same lines, an extension to the API could use a utility
class that contains only static versions of the corresponding methods
in the FastFourierTransformer class.  The original class could be
remade with only static methods, but I believe it would be far quicker
to make a FourierUtils ( :-) ) class or something.

I don't have any experience working with the junit api, but I believe
this error could be resolved by altering the programing pattern to
have the test classes in an alternate package, which would then be
able to detect this simple error at compile time.

---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org


Mime
View raw message