Re: [office] surface plots

["Andreas J. Guelzow" <aguelzow@pyrshep.ca>]

On Fri, 2009-07-31 at 15:02 +0200, Eike Rathke wrote:
> Hi Andreas,
> 
> Answer from Ingrid:
> 
> ---%<---snip---%<---
> 
> Hi,
> 
> > ----- Forwarded message from "Andreas J. Guelzow" <aguelzow@pyrshep.ca> -----
> > 
> > Date: Thu, 30 Jul 2009 15:33:02 -0600
> > 
> > On Thu, 2009-07-30 at 23:03 +0200, Eike Rathke wrote:
> > 
> >>> Gnumeric has various ways this could be specified. Gnumeric implements
> >>> chart:surface as a contour plot (not an xyz contour plot), So data would
> >>> look like:
> >>>
> >>>   A  B  C  D  E
> >>> 1 -- 1  2  3  4
> >>> 2 1  2  7  5  6
> >>> 3 2  4  3  3  2
> >>> 4 3  5  6  5  4
> >>> 5 4  3  2  3  2
> >>> 6 5  4  5  5  5
> >>>
> >>> The "y"-values would be the range B2:E6, the "x" values would be given
> >>> by B1:E1 and the "z"-values as A2:A5. So the height at x=3 z=2 would be
> >>> 6.
> >>>   
> >> The ODF 1.0 description says:
> >> "The data points are interpreted as tabular data, where each value 
> >> defines a 'height'
> >> at a specific grid location."
> >> There is nothing said, that some values are interpreted as x values, 
> >> some as y and some as height. It says *each value defines a 'height'*. 
> >> So I don't see how your interpretation could fit into that description.
> > 
> > The series would give a range of B2:E6 so all those data points are in
> > fact heights. 
> > 
> >>> The same would be true for gnumeric's surface plots.
> >>>
> >>> THe description you gave above is gnumeric's xyz contour and xyz surface
> >>> plots.
> >>>   
> >> Well nice! Come up with a proposal for a new value for attribute 
> >> chart:class to describe your xyz-surfaces and contours!
> >>>   
> >>>> Furthermore this would allow for not equidistant points 'on the grid' 
> >>>> what is not wanted for the surface chart.
> >>>>     
> >>> Why not? 
> >>>   
> >> Because this was not specified.
> > 
> > Perhaps I missed it but where were the equidistant points specified ?
> > 
> 
> The specification says that the heights should be located at a grid. 
> That is naturally understood as something equidistant. As nothing 
> further is said about the grid it must be assumed

Sorry, but there is no reason that this"must be assumed". 

>  that this grid is 
> formed by the indexes of the height values within the so called 'tabular 
> data'.
> 
> Your adding of additional values for the grid locations is in conflict 
> with this specification and is therefor wrong.

I see no conflict. Your interpretation is assumed. Apparently we are
assuming different things.

You seems to ignore the primary difference between our interpretations:
A single series with tabular data versus a sequence of series (so no
tabular data) in your interpretation.

ODF 1.1 talks about multiple series for a few chart types: scatter (this
seems to be changed for 1.2 to use chart:domain), stock, bubble (again
this was changed for 1.2 to use chart domain). By default all others use
a single series. Location in scatter and bubble in the current 1.2 draft
uses chart:domain. SO it is reasonable to assume that the same is true
for "surface".

Andreas

> 
-- 
Andreas J. Guelzow <aguelzow@pyrshep.ca>