SV: transpose on bigger datasets?

["Berntsson, Martin" <martin berntsson eon se>]

Dear Morten and all,

Very interesting discussion, and it leads me to a thought....
Isnt it time to:

* SKIP the LETTER ID of COLUMNS !!! (A,B,C,....XXX)

And replace it with numbering (1-k), perhaps with an indicator of some kind (1, 2, 3.... or c1,c2,c3... etc)?.

From my end I can see a number of advantages :
*more mathematic-like approach to a matrix
*easier to create a matrix with a given size
*Opens up for bigger matrixes'
*Referring to an element  (cell) by x(row,column).
*Give rows and columns the same "status".
*Easier programming (?)

I like what you said about 2^31 (ca 2.1x10^10), have that for both rows and columns, and let
mem set the limits. That would be very interesting.

Personally I have never liked the Letter IDing of columns, that is common in most spreadsheet softwares (MS 
Excel).
Perhaps there is a point to it for more every-day products, but for a more "high-end" software like 
Gnumerics, I think it’s the wrong choice.

OK - some ideas - I wish I could do some programming - so I could be a part of cool things come real, but 
unfortunately I don’t have that skill.
Perhaps other things...
Hope my thoughts have contributed with something.



Med vänliga hälsningar / With kind regards / Freundliche Grüße, 

Martin Berntsson