[
https://issues.apache.org/jira/browse/MATH-215?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12659564#action_12659564
]
Luc Maisonobe commented on MATH-215:
------------------------------------
I will keep the methods as they currently are.
The code is really small (about 20 lines of straigthforward code) and generics
do not add anything.
They also imply some memory overhead due to boxing and computation overhead due
to
the instanceof tests.
The fact that integer transform cannot be inverted in all cases is also a
strong indication
the implementations should stay separate.
> Fast Hadamard Transform
> -----------------------
>
> Key: MATH-215
> URL: https://issues.apache.org/jira/browse/MATH-215
> Project: Commons Math
> Issue Type: New Feature
> Affects Versions: 1.0, 1.1, 1.2
> Reporter: Daniel Kuan
> Assignee: Luc Maisonobe
> Fix For: 2.0
>
> Attachments: FastHadamardTransformer.java.diff,
> FastHadamardTransformer_generics.diff, FastHadamardTransformerTest.java.diff,
> FastHadamardTransformerTest_generics.diff
>
>
> To date, the mathematical transforms package of Commons Maths,
> org.apache.commons.math.transform, only contains implementations for the
> Fourier, Sine, and Cosine transforms.
> This issue serves to propose and track the creation of an implementation for
> the Hadamard transform.
> Definition of the hadamard transform:
> http://en.wikipedia.org/wiki/Hadamard_transform#Definition
> Unfortunately, Mathworld does not provide a very detailed definition.
> http://mathworld.wolfram.com/HadamardTransform.html
> An elegant algorithm for the fast hadamard transform can be found here:
> http://www.archive.chipcenter.com/dsp/DSP000517F1.html
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
