[
https://issues.apache.org/jira/browse/NUMBERS-65?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16443987#comment-16443987
]
Eric Barnhill commented on NUMBERS-65:
--------------------------------------
I understand the polar notation of Complex.nthRoot perfectly clearly.
I'm embarassed to say, the trig identities behind RootsOfUnity are lost in the
mists of time for me.
I popped over to Rosetta Code, every approach I saw there uses a
polar-coordinate type of approach (e.g.
[https://rosettacode.org/wiki/Roots_of_unity#Java] ). I think this is the one
to use. I agree it should be expanded to handle a negative integer as going the
other direction.
> "RootsOfUnity" vs "nthRoot"
> ---------------------------
>
> Key: NUMBERS-65
> URL: https://issues.apache.org/jira/browse/NUMBERS-65
> Project: Commons Numbers
> Issue Type: Task
> Components: complex
> Reporter: Gilles
> Priority: Major
> Labels: API, delete, gsoc2018
> Fix For: 1.0
>
>
> There is partial overlap of functionality between class {{RootsOfUnity}} and
> method {{nthRoot}} (in class {{Complex}}).
> The latter is more general (any complex number) but the former allows a
> negative argument (to specify clockwise ordering).
> Also, the implementations are different; it should be checked which is more
> accurate (and then account for possible performance trade-offs).
> API-wise, I think that we should keep {{nthRoot}} (and delete
> {{RootsOfUnity}}).
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
