[
https://issues.apache.org/jira/browse/MATH-1602?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17357108#comment-17357108
]
Alex Herbert commented on MATH-1602:
------------------------------------
I use a similar wrapper class to allow use of the classes in Commons Math 3
that require a random source.
In CM4 all these classes have either been updated to UniformRandomProvider or
moved to new projects, e.g. the distributions have moved to Commons Statistics.
With the release plan to release dependencies such as Statistics before CM4
there is not a reason to keep the adaptor. Users should upgrade to
UniformRandomProvider to use the new code.
RngAdaptor is the only class that references the interface RandomGenerator. So
it seems both these classes can be removed.
> Purpose of "RngAdaptor"
> -----------------------
>
> Key: MATH-1602
> URL: https://issues.apache.org/jira/browse/MATH-1602
> Project: Commons Math
> Issue Type: Sub-task
> Reporter: Gilles Sadowski
> Priority: Minor
>
> [{{RngAdaptor}}|https://gitbox.apache.org/repos/asf?p=commons-math.git;a=blob;f=commons-math-legacy/src/main/java/org/apache/commons/math4/legacy/random/RngAdaptor.java;h=c16587dc08162ef184fcccbbcc7b4fe9bbdcf237;hb=HEAD]
> was created to allow "skeptics" to test the then new RNG implementations in
> "Commons RNG" and compare with the RNG implementations in "Commons Math" (up
> to v3.6.1).
> Since then, the RNG codes pass stringent tests ("Diehard", "TestU01"); hence
> that rationale is moot.
> Is there another reason to keep this class?
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
