[
https://issues.apache.org/jira/browse/MATH-1124?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14163507#comment-14163507
]
Ole Ersoy commented on MATH-1124:
---------------------------------
I agree that writing a RandomGenerator mock is simple, but I do believe that in
order to maximize the simplicity, flexibility, and focus of the design it is
best to inject both the RNG and the distribution into a class/classes that
is/are designed for sampling and things related to sampling.
For instance there could be a minimal SamplingObserver interface:
subscribe(SampleListener)
next() //triggers 1 iteration of observing the sampling
next(int n) //triggers n iterations of observing sampling
I'm not an expert here, but I suspect that the strategy design pattern could be
used to design one instance that samples from all the distributions and that
could be easily extended.
> Instances of AbstractRealDistribution require a random generator.
> -----------------------------------------------------------------
>
> Key: MATH-1124
> URL: https://issues.apache.org/jira/browse/MATH-1124
> Project: Commons Math
> Issue Type: Bug
> Reporter: Ajo Fod
>
> A couple of observations:
> ... The default random generator takes a while to instantiate.
> ... Many functions of these distributions don't require a random generator.
> Generally speaking only sampling requires it.
> So, why force the default constructor to initialize with a new random
> generator ... why not use a global generic or simple generator?
> Or do away with random generator except for sampling?
> This issue was observed with the TDistribution class , but it is probably
> applicable to many classes as well.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
