Moritz,
Thanks for the info and testing. I'll quickly reply below.
Michael
______________________________
Michael Barton, Professor of Anthropology
School of Human Evolution and Social Change
Arizona State University
Tempe, AZ 85287-2402
USA
voice: 480-965-6262; fax: 480-965-7671
www: http://www.public.asu.edu/~cmbarton
From: Moritz Lennert <mlennert@club.worldonline.be>
Reply-To: <mlennert@club.worldonline.be>
Date: Wed, 11 May 2005 15:08:43 +0200 (CEST)
To: Michael Barton <michael.barton@asu.edu>
Cc: grass5 developers list <grass5@grass.itc.it>
Subject: Re: [GRASS5] New thematic mapper for GRASS 6
Michael,
On Wed, May 11, 2005 7:39, Michael Barton said:
I¹ve just finished a thematic mapper for vector maps in GRASS 6,
d.vect.thematic, and am hoping that some of you can help test it.
This is way cool !
I had been thinking about something like this for a while, but as always, I
think too much and do too little 
First superficial testing shows it works great, with just a few remarks:
- One problem I encounter sometimes (still trying to figure out the exact
sequence leading to this) is the following error:
/usr/local/grass-6.1.cvs/scripts/d.vect.thematic: line 143: TMP1: Permission
denied
I am using g.temp to generate a TMP file ID and create a TMP file to hold
data needed to calculate stats. This is actually taking place in v.univar.sh
and would be unnecessary if v.univar can be fixed/upgraded. I'm using the
format that is required for proper security in creating temporary files. I'm
guessing that there is a problem on your system in writing a file in the
appropriate locale. Is there possibly another TMP1 already created (by
another process) and in use in this area?
and then nothing is displayed as the file suppoed to contain the statistics is
empty.
See above
- The options available in the key "type" should be reduced to the only
sensible ones (point, centroid, area) as the others don't seem to work (or I
misunderstood something); also it might be useful to print a warning if the
type of thematic map display is set to "graduated points" and the vector map
type to "area" and vice versa. Currently the program runs, but nothing shows
up.
In fact, lines and boundaries should also display with graduated colors
here. So they do make sense. I haven't tested it with lines yet however.
- Concerning the legend: it would be good to have a precise indication of
where one category ends and the other starts. At the moment, it prints
something like
1.000000-3.2725 5
3.2725-5.545 7
5.545-7.8175 9
7.8175-10.09 11
The algorithm specifies the range (given above) then looks for records > 1st
number and <= 2nd number. I understand what you are wanting to do, but not
sure how to do it. I'll see if I can figure out something as I refine this.
so if you have a value of 3.2725, you don't know in which category it is...
If I understand the code correctly it is in the first category and the legend
should therefore be:
1.000000-3.2725 5
3.2726-5.545 7
or some other way of saying that the range goes from minimum (exclusive) to
maximum (inclusive).
BTW the way, this raises the question of what happens if a feature has exactly
the $min value of the first category ?
Right now it gets left out. But this I can change (I think).
And just to prove once again that users are a terribly ungrateful bunch and
always want more, here a few wishes:
- I think one useful addition would be the possibility to display the
available columns, or, even better, to chose from a list of available columns.
I'd love to do this, but don't know how to program this in the existing
g.parser interface via shell scripting at the moment.
- It was very easy for me to very quickly create a new color scheme (here we
use blue for low values and red for high values...) in the script. That's very
useful for those of us who are not afraid of going into the code. But it would
be interesting to see if there is a way to offer a color choser to chose the
min and max color and to automatically calculate the intermediate values.
Another great option would be to also be able to chose all the colors
manually...
I'd also like a color chooser (kind of like in r.colors rules). I have an
idea of how this might be accomplished. I don't know if this would make it
harder (i.e., too complicated) or easier for most people to do.
I thought about a manual color chooser, but realized that by making a group
in the GIS Manager and duplicating a vector display line, there is already a
much better interface for doing this than I could do in a script.
Congrats Michael, this will make GRASS much usable for all the vector bunch
out there !
Thanks for your input. It is indeed appreciated. I'm glad if it helps those
who are using vectors.
Now if someone could just figure out how to re-write d.legend for vectors
also....
Michael
Moritz
