Re: Median: Oasis and Fast Sorting Algorithm

["Andreas J. Guelzow" <aguelzow taliesin ca>]

On Sat, 2007-10-02 at 19:35 +0200, Leonard Mada wrote:
> Lets assume the following space A={1,2,3} [I do not know if it is called 
> space in English, probably NOT.]
>
> Lets take the array 1,1,2,3,3,3
>
> So the median is (2+3)/2 = 2.5 ? BUT the space A does NOT have the value 
> 2.5. It is either 2 or 3. In this case, it is more appropriate to talk 
> about 2 medians, 2 and 3! And this list DOES NOT have the middle value 
> of 2.5 because this value DOES NOT exist.

I am pretty sure that at least 99% of the spreadsheet users don't care
about your space {1,2,3}. FOr them 2.5 exists.

> Weighted Medians: actually, most real world applications making use of 
> median from signal processing, use the weighted median. (Just google for 
> "weighted median".) So again, because Excel did take the easy ride, I do 
> NOT believe that a standard should follow the same way.

Excel, OOoCalc, Gnumeric, KSpread,... agree on what MEDIAN should
alculate. The OpenFormula standard is not intended to invent new
functions but to standardize these functions to allow for consistent
interchange using the odf format.

Perhaps there should be another function MEDIAN_MADA which would clearly
need more arguments to allow for a specification of the space in
question. 

Andreas Guelzow
>


-- 
Prof. Dr. Andreas J. Guelzow
Dept. of Mathematical & Computing Sciences
Concordia University College of Alberta