[Kst] [Bug 111239] Ability to easily setup supressed axis plots

[Barth Netterfield]

On Wednesday 21 June 2006 20:43, Andrew Walker wrote:
> ------- I think I see what you're trying to achieve, but the reasoning is
> not clear to me:

sorry.... here we go....

> why would you want to merge axes on plots that are not touching?

we don't - but if some one has arranged them by hand, and is off by a few 
pixels, we want to be a little generous... maybe 5% of the width of the plot? 
Maybe less is ok too.

> why would you want to merge axes on plots that do not have the same
> horizontal extent? It would seem simple enough for the user to cleanup
> layout to ensure they do.
Unless the user has some plots wider than others.  We do this.  So as long as 
they are 'close' we should consider them the same.
>
> what if plots are overlapping vertically? Which axes are hidden?

by a few pixels? accept it - its probably just imprecise plot positioning.  By 
a lot?  Don't accept it - they are doing somehting else.

> What if you have a series of, say 10, plots which are left-aligned in a
> single column but whose widths increase by 5% with each plot? Which ones
> are included in the codensation?

Maybe 5% is too generous.  On the other hand, if they were selected, we should 
assume it was for a reason.  If there is a plausable condensation partner, 
then we should condense.  

The biggest problem is: 
Make 10 plots, in three columns with the datawizard.
In layout mode, select all 10, and condense X axis.  Only the first 3 rows get 
condensed.  I would expect that all of them would condense, so there would 
only be axis numbers on the bottom plot of each of the three columns.  

Or

Make a grid of plots: select one column.  Condense X axis.  Nothing happens.  
The column we selected should be condensed.

So, the rule should be, if there is a plausable selected plot to condense 
with, condense with it.

cbn