[RkWard-devel] Distribution plots done. (testing needed)

[Thomas Friedrichsmeier]

Hi!

On Monday 05 February 2007 08:58, Prasenjit Kapat wrote:
> Finally I have completed my work on the distribution plots. Modifications
> in this version are:

Great! Thanks a lot.

I have committed everything (with some small changes, described below), but 
testing will take some time.

> [3] plot_*_distribution.rkh: Help files for all the distributions created.
> Detailed description given for "Plot Beta distribution"  and the rest
> linked to it. Please read the Beta help file and comment.

Looks good to me. You might consider removing the
	The default values should produce a plot of the "standard" distribution.
On the very first look, I thought "Hm, but this was supposed to be the beta 
distribution, no? And what's the "standard" distribution, anyway?" Until I 
realized, what it's supposed to mean. I think users in want of the "standard" 
will simply leave all settings untouched, and try whether it gives the 
desired result. But what does everybody else think about this?

> 1. plot_options.xml: xlim/ylim not restricted to be real.

True. The alternative would a spinbox with an additional checkbox for "use 
default". However, that would make changing the limits quite a hazzle. Of 
course a nice side-effect is that you could also use R symbol names (like 
e.g. "pi").

> 2. plot_options.xml: checkboxes for "interpret as R expression" not
> initially disabled for main/sub title, like it is done for xlabel and
> ylabel.

Ok, I added this, for consistency with xlab/ylab.

> 3. plot_options.{php,xml}: The tab character is interpreted literally in
> the input fields (like xlab/ylab/xlim/ylim etc.) instead of changing the
> focus to the next field.

Pretty annoying, indeed. I'll have a look later.

> 4. plot_options.php: Interpret as R expression is not quite transparent. It
> would be better to use, expression() in the php code itself. As of now,
> expression(..) needs to be literally typed into the input box.

The current meaning is this: unchecked -> quote the given string; checked -> 
do not quote (and hence let R parse this). Maybe the label "expression" is a 
bit misleading. Do you have a suggestion?

> 5. Should closing the parent plugin close the embeded plugins too? As of
> now, such is not the case.

Fixed.

Two changes I have done to your changes:
- Added an external property "default_pointtype" to plot_options, and move 
the "$type_tag" logic into the plot_options plugin. Adjusted the discrete 
plots accordingly.
- Remove all "max_precision"-attributes from the spinboxes. This parameter 
used to limit the maximum value selectable (and the real number had to be 
mapped to an integer divided by a constant), but it's mostly obsolete in the 
upcoming version. It should no longer be used, unless there is a good reason 
to limit the maximum precision the user can enter.

Regards
Thomas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://mail.kde.org/pipermail/rkward-devel/attachments/20070205/75475dc2/attachment.sig>