Thanks for the encouragement Markus.
To make these scripts 5.7 ready, what is mainly needed is to make sure they have command line options (as opposed to being interactive) and to add the tcltk dialog support to the header.
Given this, it seems to make more sense to simply use these scripts in 5.3 and slightly rewrite the 5.3 GUI to recognize this. The only thing that is missing is a usage statement when you type -h from the command line. g.parser generates this automatically in 5.7 but not in 5.3. However, doesn't seem to like it if I script it (as has been done previously). I'll mess around a bit with finding a way to generate a usage statement that works for both systems, modify the 5.3 GUI appropriately, and I think we can use the same scripts in both. You can link either way that way (5.3 to 5.7 or 5.7 to 5.3), which ever works best for you.
That said, I have a thought about a new protocol for scripts for GRASS 6 that I wanted to run by you.
1. All scripts integral to GRASS (e.g., d.m) should go in $GISBASE/etc or be transformed into C++ modules.
2. All other scripts would be considered optional 'user scripts' and installed at run time from whatever is in the $GISBASE/scripts directory.
3. All user scripts must have a header for an autogenerating tcltk dialog (or something equivalent for an alternative GUI standard) as now, but with an additional 'menu' section. the menu section has a line to indicate where in the menu hierarchy the script will install and a line indicating what its menu name will be. [Alternatively, all scripts could go under a separate 'user script' menu item].
4. At run-time, GRASS would read the files in the $GISBASE/scripts directory and automatically install them into the menu, based on their headers. Scripts with incorrect menus will simply not load into the menu and not generate an error message (although they could be callable from the command line if they work otherwise).
This pretty much a 'plug-in' protocol. If such a protocol could be developed and promulgated, there could be a script download repository for all scripts and they would no longer be distributed with the base GRASS distribution. They would be user installable options, and people would be encouraged to submit scripts to the repository. This is kind of like the ImageJ model
What do you think?
Michael
On Wednesday, April 14, 2004, at 07:19 AM, Markus Neteler wrote:
Michael,
thanks for your efforts!
May I suggest that we update most scripts in 5.3 and
simply link into 5.7? To have two versions increases
the workload.If you find someone to update 5.3 scripts with your
versions, I'll link the stuff properly into 5.7.Only if they really differ we'll have to maintain two
scripts (5.3/5.7).Markus
______________________________
Michael Barton, Professor & Curator
Department of Anthropology
Arizona State University
Tempe, AZ 85287-2402
USA
voice: 480-965-6262; fax: 480-965-7671
____________________
C. Michael Barton, Professor
Department of Anthropology
PO Box 872402
Arizona State University
Tempe, AZ 85287-2402
USA
Phone: 480-965-6262
Fax: 480-965-7671