Re: removing zscore and strtrim that were moved to core

[Ben Abbott]

On Jan 19, 2012, at 12:21 PM, Carnë Draug wrote:

> On 19 January 2012 17:07, Arno Onken <address@hidden> wrote:
> 
>> On 01/19/2012 05:38 PM, Carnė Draug wrote:
>>> 
>>> I have just removed the strtrim funtion from the string package.
>>> 
>>> I want to do the same for the zscore function from the statistics
>>> package. Is there any reason not to so?
>> 
>> Removing the zscore function from the statistics package might break legacy
>> code. Compared to the Octave core zscore function, the zscore in the
>> statistics package also returns the mean and standard deviation.
>> 
>> The best way to deal with this would probably be to change the Octave core
>> function to also return mean and standard deviation and then to remove the
>> zscore function from the statistics package. However, the core zscore
>> function uses the center function instead of calculating the
>> mean directly. So changing the core zscore function would complicate it
>> considerably. I am not sure if this is desirable and what's the best
>> solution.
> 
> It seems to me that matlab zscore also returns the mean and standard
> deviation so that's something that should be added to the core
> function. Could zscore in core be changed to do it in in one or the
> other way depending on the value of nargout? (I'mm CC'ing the
> mantainers mailing list)
> 
> Carnë

I hadn't realized there was a zscore function. The polyfit() function should 
probably call that instead of doing the normalization itself.

I don't have a copy of the ML statistics toolbox. I notice that Octave's 
implementation returns a result whos class matches that of the input. Does ML's 
behave that way as well ? (the on-line docs don't mention it).

Ben