Re: sec : U Re: MPI_Plot Was: Something else

[davidf(at)dfanning.com (David Fanning)]

Dave Greenwood (greenwoodde@ornl.gov) writes:

> Well - almost anyway ;-)  Actually, I think the problem is more likely
> to be related to my 8-bit display than to VMS.  When I click on the
> axis color box, I get a complaint about a missing SADDLE BROWN color
> and the axes turn white (ie disappear).  When I select a new color,
> the axis reappear with that color but all 3 color boxes in the "Plot
> options" widget turn the axis color.  I can continue to manipulate the
> colors independently, except that the 3 boxes always show the axis color.
> 
> If I say "mpi_plot,color='green'", then when I click on the modify button
> the axes change from green to the data color (navy by default).  If I
> then change the data color, the axes (and color boxes) revert back to
> green.
> 
> Nevertheless, "nifty" seems an appropriate comment.

Yes, colors are a bit of a problem on 8-bit displays,
for a couple of reasons. One was just boneheaded programming,
and forgetting to check everything on an 8-bit display.
(Your dear programmer had just assumed all 8-bit displays
had been sold to third-world countries years ago. He
completely forgot about government agencies.) Anyway,
*that* problem has been solved in the version I put
up just minutes ago.

The other problem is more fundamental. This program
combines both direct and object graphics, which is
a disaster on 8-bit displays. The problem is that the
one hardware color table is used in two completely
different ways by the two graphics systems. If you have
a direct graphics window open and you put up a window
with object graphics in it, the direct graphics window
goes crazy, and visa versa. Of course, you don't see
any of this on a 24-bit display, since there window
colors are not directly connected to the colors in
the hardware color table. 

Two years ago, I wouldn't have even thought about
writing a program that combined direct and object
graphics. These days, you can *almost* get away
with it. A year from now, I doubt it will even be
an issue.

The other problem also has to do with changing colors
in color displays. My Pickcolorname and FSC_Color 
programs normally have 88 colors they know by "name".
But, these colors often need to be loaded into the
color table to "do things" with them. Even though
I'm only mucking with the color table temporarily,
and I plan to restore it, this will cause graphics
windows that use those 88 color indices to get a little
funny. So, on 8-bit displays I use only 16 colors
in these two programs, hoping to limit the damage.
Unfortunately, I had picked default colors for MPI_PLOT
that were not among these 16 (duh!). That's fixed
now, too.

Many thanks for all the comments. I think it is 
getting there. :-)

Cheers,

David

-- 
David Fanning, Ph.D.
Fanning Software Consulting
Phone: 970-221-0438 E-Mail: davidf@dfanning.com
Coyote's Guide to IDL Programming: http://www.dfanning.com/
Toll-Free IDL Book Orders: 1-888-461-0155