Spatial Modeling Environment

[Steffen Meyer]

The group I work with is looking to develop a general 
interface into a number of spatial models that will be written 
for siting purposes.  

I am looking for approaches/tools to acheive a general interface to 
the user modification of desired values for ranges of values 
in r.mapcalc, r.infer, r.combine, and r.weight script.
I envision an interface that requires a general GRASS spatial 
model to have a set of start-up values for various portions 
of the model execution that pursues siting based on user preferences
and user weights.  The interface would substitute appropriate values 
into the GRASS script, and then run the script to develop the 
outcome map(s) and review the results visually and numerically.

I anticipate this effort would require adopting tools or code
previously developed if possible.  I am trying to not reinvent any
wheel(s), and would like to capitalize on work that has not received 
extensive press in the ftp documents.  The preferred interface would 
support X/Motif similar to XGRASS and the display monitors currently
supported in GRASS4.x.

Ideally I would gain access to a graphical tree tool that 
could support an expanding/collapsing tree.  Each segment of the tree
must support 1 or more maps feeding into a single operation that 
results in 1 or more maps.  
  If the user would click on a map 
    call-backs would launch desired display routines
  If the usr clicks on the operation 
    the input maps, the output maps and the parameter sequence 
    would be revealed for each type of operation

A simpler interface will read the GRASS script spatial analysis 
model and permit changing the input for each operation in the 
model from top to bottom.

Are any you GRASS programmer types aware of any efforts in this area?
I have looked at the code to xgrass and ruled out some of the 
functionality provided there.
Does any have any experience with 
an X version of parameter modification/substitution.  

We have accomplished a similar interface using a public domain 
stick editor tool, but the graphical representatin is not very 
clean.  Essentially this is a concept for a generalized spatial 
model interface.

Thanks for you thoughts
Steffen Meyer
Integrated Decision Support Group
Colorado State University