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] (MATH-1137) BOBYQA incorrect indexing
Date Fri, 11 Jul 2014 17:04:05 GMT

    [ https://issues.apache.org/jira/browse/MATH-1137?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14059002#comment-14059002
] 

Gilles commented on MATH-1137:
------------------------------

bq. The purpose of testing is to find errors, not to pass tests.

If there is no test to leave a trace of something that has been changed, then that thing can
be changed again later without anyone noticing.

bq. Do you have any idea where this code is described in the BOBYQA paper?

No.
One goal of CM is that the _implementation_ should be documented (or self-documenting).
IMO, we'd be better off if the algorithm were reimplemented from scratch (as was already discussed
on the MATH-621 page).

bq. I can verify I have run BOBYQA hundreds of times on lots of problems, but have no idea
whether this bit of code has ever been exercised.

That would be quite easy, if you can easily rerun those codes. You'd just have to uncomment
line 538, and the code will throw an exception if it ever enters the block.


> BOBYQA incorrect indexing
> -------------------------
>
>                 Key: MATH-1137
>                 URL: https://issues.apache.org/jira/browse/MATH-1137
>             Project: Commons Math
>          Issue Type: Bug
>    Affects Versions: 3.3
>            Reporter: Nigel Goodwin
>




--
This message was sent by Atlassian JIRA
(v6.2#6252)

Mime
View raw message