Re: rudimentary tick positions

[Michael Goffioul]

On Tue, Feb 5, 2008 at 6:39 AM, Shai Ayal <address@hidden> wrote:
> Well, we already have get_axis_limits in graphics.cc which attempts to
> find the "nicest" limits, So the limits are probably "nice" by the
> time we get to calc_ticks. However I will look at them to see that
> they are "compatible" in the sense that they both agree on the meaning
> of "nice"

The auto-scaling should be aware of the current tick spacing, otherwise
you won't get nice results. For instance, look at the following screenshot
obtained with "plot(rand(1,100))". As you can see below, the xlim should
have been extended to [0,100] and ylim to [0,1]. OTOH, if limits mode
were manual, then the tick arrays should not have contained out-of-bounds
ticks (xticks/yticks should not contain 0 and yticks should not contain 1).

octave.exe:10> get(gca, 'xlim')
ans =

     1   100

octave.exe:11> get(gca, 'ylim')
ans =

   0.020000   0.980000

octave.exe:12> get(gca, 'xtick')
ans =

     0    20    40    60    80   100

octave.exe:13> get(gca, 'ytick')
ans =

   0.00000   0.20000   0.40000   0.60000   0.80000   1.00000

screenshot.gif
Description: GIF image