20010223: graphic line width/dash pattern; proftomd decoding bug

>From: "Paul L. Sirvatka" <address@hidden>
>Organization: College of DuPage
>Keywords: 200102162013.f1GKD9L01250 McIDAS-X SFCCON GRDDISP

Paul,

>I hope skiing was good and that you are intact today!

Skiing was fantastic, and I am still in one piece, so it was a good day :-)

>I am getting the data thing! At least a good deal of it. Therefore many
>things are good.

Excellent.  My view of setting up ADDE datasets is that once you do
a couple, the process makes a lot of sense and is easy to do.

>2 quick things...
>
>How do you change the dash length? RAOBCON is more of a dot than a dash
>comapred to SFCCON.

Hmm...  Since I never notice things like this, I decided to do a series
of RAOBCON and SFCCON plots:

SF 1
RAOBCON Z 500 NACONF 12 DASH=ALL
SF 2
SFCCON PRE NACONF 12 DASH=ALL
SF 3
RAOBCON Z 500 CO DASH=ALL PRO=CONF
SF 4
SFCCON T CO 12 DASH=ALL PRO=CONF

As can be seen by the series of GIF (tm) images, all of the dash lengths
were the same:

http://www.unidata.ucar.edu/staff/tom/gifs/raobcon1.gif
                                      ...  sfccon1.gif
                                      ...  raobcon2.gif
                                      ...  sfccon2.gif

I don't doubt that the dash patterns can be the same, however, since their
form is governed by the global dash pattern settings that are controlled
by the GD command:

GD -- Sets graphics display parameters
  GD INI
  GD width dlength gcolor glength
Parameters:
  INI     | initializes the graphics parameters to their logon values
  width   | graphics line width in pixels; range 1-64 (def=last value set;
            at logon it is set to 1)
  dlength | pixel length of dashes in lines; range 0-64 (def=last value
            set; at logon it is set to 10)
  gcolor  | graphics color level of gaps in dashed lines (def=last
            value set; at logon it is set to 255, transparent)
  glength | pixel length of gaps in dashed lines; range 0-64 (def=last
            value set; at logon it is set to 10)
Remark:
  Enter GD with no parameters to list the current graphics settings.
Example:
  GD 3 5 3 5
  Sets the line width to 3 pixels; sets the length of dashes and gaps in
  the dashed lines to 5 pixels; sets the color of the gaps to color
  level 3.
----------

I have always argued that setting the dash line pattern globally is
_bad_.  It should be settable on an application-by-application basis.
The problem with setting it globally is that the pattern set by a GD
invocation remains in effect until one resets the pattern with another
GD invocation.  This will have unintended effects, for example, after
setting the line width to more than one and then drawing something that
should have a line width of 1.

>Also...
>
>I have some problems plotting out 6min profilers. I think the data is
>there...it plots out the ID's, but when I do a timeplot, it *usually*
>cannot find any data, although it did for one of them.

Hmm...  I am wondering if this is the same sort of problem that someone
reported some time ago.  I found that there is a bug in the profiler
decoder, proftomd, that incorrectly sets the number of stations that
have been decoded in an MD file.  As I recall, if one tries to plot
wind profiler data for certain stations, they would never plot while
other stations always plot.  As I recall, the stations that had
problems in the plot were those in IL.  The reason for this is that
they are at the end of the list of stations that are decoded.  So, the
fix would be to upgrade your profiler decoder to one with the corrected
code.  I do not have a new ldm-mcidas decoder release with this in it
(it is just me, and I am totally swamped), but I can get you going if
this is your problem.  Please let me know.

>Is there something I do not know about setting up the plots?

I doubt it.

>I will be more specific if need be.

No, I think that you are running into the same problem.

Did whoever installed the ldm-mcidas set of decoders on your machine
build them from scratch, or did s/he install them from binaries?  If
s/he built them from source, I can stick the corrected proftomd.c
source file out in the pub/ldm-mcidas directory for you to grab.  The
decoder would have to be rebuilt and reinstalled.

If the installation there was from binaries, then I can stick an new
proftomd binary for Linux 2.2.x kernels out in the
pub/binary/linux_2.2.x-i686 directory.

Please let me know which way to go.

>Thanks!

Later...

Tom