Hi Jordi,
if you believe, that I'm doing the NaN-tb because of a petty war you are
grossly mistaken. The NaN-toolbox tries to solve a real issue - and it
does it very well, I think. I also do not understand your issue - You do
not need to use the NaN-toolbox if you do not like it. So what is your
issue?
Concerning your question: NA-skipping instead of NaN-tb is not a
solution, at least not for the NaN-toolbox for the following reason:
o) When you compute in statistics some expectation value, it does not
matter whether there is a NA or a NAN, both should be skipped.
This is the main reason why the NaN-toolbox will not move to skipping
only NA's. There are also some other reasons why NA-skipping instead of
NaN-skipping is a bad idea.
- NA do not make things simpler but more complicated. There are no clear
rules when NA and when NAN's should be used. NaN's can be confusing, so
introducing NA in addition make it even more complicated.
- Having a nanmean() and a namean() is not simpler than having nanmean()
and mean(). Eventuelly, you might even need three versions that do the
same thing, except for the NaN- and NA-handling behavior.
- NA can cause a significant performance penalty. ISNAN() is supported
by hardware, but ISNA() needs to analyze the payload of NaN which is
much more complicated.
Some final remarks on NA. Nobody is using it, and I really do not see
any advantage of NA. If NA's would provide a solution, why do the
statistical core functions of Octave not use it? I assume there is a
reason for it, and the reason is that NA's do not provide any real
advantage.
Alois
On 02/28/12 23:42, Jordi Gutiérrez Hermoso wrote:
> Ping on the discussion below. I really would like to settle this once
> and for all and stop this petty war we've been having about shadowing
> core functions and then enabling warnings when core functions are
> shadowed.
>
> I think skipping NA instead of NaN would be desirable and a good agreement.
>
> Alois?
>
> - Jordi G. H.
>
> 2012/2/21 Jordi Gutiérrez Hermoso<jord...@octave.org>:
>> On 21 February 2012 02:48, Dr. Alexander Klein
>> <alexander.kl...@math.uni-giessen.de> wrote:
>>> Am 20.02.2012 um 17:23 schrieb Jordi Gutiérrez Hermoso:
>>>
>>>> I'm not sure what the correct solution is here, but perhas we
>>>> should implement NA-skipping behaviour in the functions that Alois
>>>> overwrote.
>>>
>>> Hmmm ... these are pretty many. How would it affect performance on
>>> in-order architectures like Intel's Atom if for example sum had to
>>> check every summand for being NA?
>>
>> I don't know, it might not be that much. Whatever the performance
>> penalty is, it seems to be acceptable to Alois and users of his NaN
>> package.
>>
>> I'm trying to offer some reconciliation here. Since Alois insists that
>> Octave is wrong for not skipping NaNs, perhaps he would be satisfied
>> if it skipped a special NaN bit pattern instead?
>>
>> Oh, I just discovered that this is an acient discussion. Perhaps it's
>> time to revive it?
>>
>>
>> http://octave.1599824.n4.nabble.com/Skipping-NA-s-in-statistical-functions-td1654304.html
>>
>> - Jordi G. H.
>
> ------------------------------------------------------------------------------
> Keep Your Developer Skills Current with LearnDevNow!
> The most comprehensive online learning library for Microsoft developers
> is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
> Metro Style Apps, more. Free future releases when you subscribe now!
> http://p.sf.net/sfu/learndevnow-d2d
> _______________________________________________
> Octave-dev mailing list
> Octave-dev@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/octave-dev
------------------------------------------------------------------------------
Virtualization & Cloud Management Using Capacity Planning
Cloud computing makes use of virtualization - but cloud computing
also focuses on allowing computing to be delivered as a service.
http://www.accelacomm.com/jaw/sfnl/114/51521223/
_______________________________________________
Octave-dev mailing list
Octave-dev@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/octave-dev
