Re: Merging Octave and Octave-Forge?

[Jaroslav Hajek]

On Wed, Aug 27, 2008 at 9:53 AM, Thomas Weber
<address@hidden> wrote:
> Am Dienstag, den 26.08.2008, 17:23 -0400 schrieb John W. Eaton:
>> On 26-Aug-2008, Levente Torok wrote:
>>
>> | Perheps because Function cov() works differently on matlab and on
>> | octave,
>>
>> In what way?  Has this been discussed before?
>
> They provide a package where NaN values are ignored during calculations.
> cov.m is one place where they needed to change the code. In case where
> cov in Octave and Matlab differed, they chose the Octave way
>
> No, I don't use the package, but it's clearly explained in the README
> and the code for cov.m.
>
>
>> | Alois Schloegl and Matthew W. Roberts replaced the octave version with its 
>> own version in
>> | the extra/nan-1.0.6 package.
>> | In the installation README file, the authors suggests us to completely 
>> overwrite
>> | the original /m/statistics/base/*.m,*.oct files.
>> | This can be seen as an obstructive and collabortive act on behalf of 
>> octave-forgers.
>>
>> I think we should discourage people writing packages from replacing
>> core functions.  But I don't see that we can stop them from doing it.
>
> I disagree. It's quite obvious they had a need for treating NaNs as
> missing values instead of usual calculation rules.
>
> If you don't want that behaviour, don't install the package. The
> package's documentation is totally clear about the package's effects.
>

Why didn't they follow the "convention" of Matlab's stat toolbox, and create
nancov, nancor etc? Given that there is nansum, nanstd, nanmean...
That way, both kinds of functions could be used at once.


>
>> Perhaps Octave should warn about duplicate functions in the load-path?
>
> Why? Sorry, but if people install stuff blindly, the only sensible
> resolution is shipping brain-1.0.0.
>
> And I routinely copy functions from Octave into the current directory to
> change them (easist way for bug finding and fixing, either in my or
> Octave's code.)
>
>        Thomas
>
>



-- 
RNDr. Jaroslav Hajek
computing expert
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz