On 2011-07-03 19:16+0100 Steve Schwartz wrote:

> Alan,
>
> On Fri, 2011-07-01 at 18:11 -0700, Alan W. Irwin wrote:
>> The attached plot (generated using test_style.py and -dev pngqt) shows
>> both the new and old patterns.
>
> Out of curiosity, I checked to see what we do in qsas. We "roll our own"
> linestyles using plstyl to offer our users 5 linestyles:
>
> solid, dashed, dash-dot, dotted, and dash-dot-dot-dot
>
> and give them a parameter to scale (stretch) the pattern.

Good idea.  Half my changes for the new line styles theme involved
simply stretching everything by 1.5, but other users would want to try
different stretch factors.  Therefore, we should implement a
user-selectable stretch factor for PLplot similar to what QSAS does
in addition to implementing convenient line style themes.

> I attach a
> sampler, which uses a factor of 2 for the stretch factor. These are
> pretty similar to your new styles 1,3 ,5 ,2 and 8 although I think our
> "dots" are shorter than yours.

I struggled with the "short dots" question myself when putting
together my first cut at a new line styles theme. The issues are it
will only be a dot (or more exactly a square) for just one particular
line width, and if you try to use too short a length for a dash, the
PLplot internal routine that plots the dashed line skips that short
dash completely.  (So this would be a limitation on how far you could
compress the result using a stretch factor of less than unity.) In the
event, I ended up using rather long dashes for my "dots" as you
noticed.  The result is good enough for the research plot I just
finished.  I will revisit this issue for my second try at such a
theme, but I will have to put that off until much later because of my
time constraints this summer.

>
> I'm inclined to agree that the old plplot linestyles are less than
> optimal. I don't have strong views on how plplot might move toward
> better default line styles. Since we do our own anyway, your Option 1
> wouldn't affect us at all. But while some users are bound to rejoice and
> say "at last" others will be annoyed that they don't get the same plot
> tomorrow that they did yesterday because someone changed something on
> them.

Exactly.  So I think we should probably stick to the current less than
optimal default, but make it convenient to change from that default to
something else. That is why I think the way to go is to implement new
functionality that allows you to update the mark and space arrays
corresponding to any given pllsty index or update all the pllsty
choices at once by reading mark and space values from a "line-style"
file.  In other words follow the same ideas we use to allow
user-chosen themes for cmap0 and cmap1 (with the default theme at
run-time being what it has always been unless you go out of your way
to configure a different run-time default at the CMake stage).

I have just created a feature request at SF for PLplot concerning line
style themes so we don't lose these ideas.

Alan
__________________________
Alan W. Irwin

Astronomical research affiliation with Department of Physics and Astronomy,
University of Victoria (astrowww.phys.uvic.ca).

Programming affiliations with the FreeEOS equation-of-state implementation
for stellar interiors (freeeos.sf.net); PLplot scientific plotting software
package (plplot.org); the libLASi project (unifont.org/lasi); the Loads of
Linux Links project (loll.sf.net); and the Linux Brochure Project
(lbproject.sf.net).
__________________________

Linux-powered Science
__________________________

------------------------------------------------------------------------------
All of the data generated in your IT infrastructure is seriously valuable.
Why? It contains a definitive record of application performance, security 
threats, fraudulent activity, and more. Splunk takes this data and makes 
sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-d2d-c2
_______________________________________________
Plplot-devel mailing list
Plplot-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/plplot-devel

Reply via email to