<html>
<head>
<base href="http://bugzilla.scilab.org/" />
</head>
<body><br>
---------- Bug Summary ----------- <br>scatter fails with 1 data point <br><br>
<br>
<p>
<div>
<b><a class="bz_bug_link
bz_status_NEW "
title="NEW - scatter fails with 1 data point"
href="http://bugzilla.scilab.org/show_bug.cgi?id=15481#c3">Comment # 3</a>
on <a class="bz_bug_link
bz_status_NEW "
title="NEW - scatter fails with 1 data point"
href="http://bugzilla.scilab.org/show_bug.cgi?id=15481">bug 15481</a>
from <span class="vcard"><a class="email" href="mailto:sgougeon@free.fr" title="Samuel GOUGEON <sgougeon@free.fr>"> <span class="fn">Samuel GOUGEON</span></a>
</span></b>
<pre>(In reply to Samuel GOUGEON from <a href="show_bug.cgi?id=15481#c2">comment #2</a>)
<span class="quote">> a) The parsing and tests of their input arguments is not robust. See your report. Another trivial example of inappropriate message:
> --> scatter(1:4, 1:4, 30, "#ff5566");
> ATTENTION : Incorrect number of inputs for property-value pairs.</span >
Actually, this message is presently correct. I was misled because some allowed syntaxes are not documented.
But other syntaxes actually generate inappropriate messages.
(In reply to Oscar van der Velde from <a href="show_bug.cgi?id=15481#c1">comment #1</a>)
<span class="quote">> I must note that scatter does not seem to have an option to specify the data value range for the marker size and color. This becomes a problem
> when trying to use scatter as I do, plotting real data of which the points are usually a selection from a wider range of possible values (imagine
> for example lightning lat/lon locations with markers sized by peak current).
> So the range in my 1-point case cannot be properly determined from the data min/max itself.</span >
Ok, i think i have understood your point. In a forthcoming Scilab Enhancement Proposal (SEP), i will propose to add a <"mapColors",colorsLUT> pair of
options, as well as <"mapSizes",sizesLUT>, that will work with msizes and mcolors data:
* min(mcolors) will be drawn with colorsLUT(1)
* max(mcolors) will be drawn with colorsLUT($)
* any mcolors element will be drawn with an interpolated color between colorsLUT(1) and colorsLUT($).
* colorsLUT could have intermediate values, for piecewise linear interpolations.
* It will be possible to specify colors with [r g b] intensities, or their "name", or their "#RRGGBB" hexa codes, or their indices in the colormap.
Same thing for <"mapSizes",sizesLUT>
Is it what you had in mind?
Other options will be proposed.
The SEP will be posted on users@ for discussion.</pre>
</div>
</p>
<hr>
<span>You are receiving this mail because:</span>
<ul>
<li>You are watching all bug changes.</li>
</ul>
</body>
</html>