If I select a block of data with zeros across the top row, the guppi plots
end up empty.

1) new gnumeric (1.0.3-cvs)
2) B4:E7 gets 
0    0    0    0
12   23   32   34
23   32   12   23
45   43   34   24

3) select B4:E7
4) Plot either line with points or scatter
-> plots end up empty

In a line plot, Excel assumes that we are plotting four series (the rows)
and assigns x values of 1,2,3,4 to each column. The first series is a line
on top of the x-axis. In an X-Y scatter plot, excel makes the point markers
stradle the y-axis at x=0 and extends the x-axis to x=1. 

I'm not quite sure what we would want to do since Excel has two kinds of
line plots (those on their own and those withing "X-Y (scatter)" We might
want to mimic Excel when the user simply selects a block/line plot (i.e.
add spurious x values of 1,2,3,...num_cols_in_selection). Where data all
have identical X values perhaps the best would be to center the data and
have an x-axis only labeled under the points (and non zoomable?).

Note that gnumeric's response to the *way* in which the selection is made
is way cooler than excel who ignores a user who selects by column or by row.

a bit confused,
adrian