commons-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Gilles (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (RNG-32) Implement more generators
Date Mon, 01 Apr 2019 22:09:00 GMT

    [ https://issues.apache.org/jira/browse/RNG-32?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16807232#comment-16807232
] 

Gilles commented on RNG-32:
---------------------------

bq. Do you think working on this issue requires deep mathematical knowledge?

No (we "only" port/implement existing algorithms; we do not try to design new RNGs).
Unit tests are required to prove that the port is correct.

See the "dev" ML archive for a [recent discussion|https://markmail.org/message/4cxdgtufj5chsp3m]
with a GSoC candidate.


> Implement more generators
> -------------------------
>
>                 Key: RNG-32
>                 URL: https://issues.apache.org/jira/browse/RNG-32
>             Project: Commons RNG
>          Issue Type: Wish
>            Reporter: Gilles
>            Priority: Minor
>              Labels: contributors, gsoc2019, scope
>         Attachments: lsf.java
>
>
> Commons RNG is focused on pure-Java implementations of standard deterministic generators.
> Quite a few algorithms could be added, but priority is on fast generators that generate
sequences of _pseudo-random_ numbers; i.e. the requirement is strong _uniformity_, but *not*
strong _unpredictability_ (a.k.a. _true_ random numbers).
> In particular, in Commons RNG, there is no provision for using an external entropy pool.
> Beware that some well-known (and much used) algorithms have been proven to fail spectacularly
on the uniformity requirement.
> Would-be contributors should look at the {{commons-rng-core}} module for how to implement
a generator, and at the {{commons-rng-examples}} module for how to test the uniformity requirement.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Mime
View raw message