[GRASSLIST:2062] Re: GRASS v ESRI

[Eric G. Miller]

On Thu, Jul 05, 2001 at 04:54:50PM +1000, Richard Ford wrote:
> Who has used Mapserver.
> 
> I guess I am now learning the difference between a "Full powered" desktop
> GIS.  And 'other' solutions that take GIS data and put them on the web with
> some limited GIS interaction features.
> 
> Again, any full on GIS expert, GRASS experts have an opinion?

Not to dissuade you from using GRASS, but my humble opinion is it isn't
well suited for internet mapping applications.  You'd have to put a fair
amount of effort into session management.  Also, I wouldn't necessarily
trust GRASS to be safe from malicious hackers -- so guard how you take
input and put the system together.  Since you wrote about updating data
via the web, I presume you already anticipated some kind of user
authentication/session management protocal.

GRASS probably is most analogous to Arc/Info, though in some ways is
less functional (vector stuff) and in other ways is more functional
(raster/imagery).  It's primarily an analysis environment, and primarily
raster based.

FYI: A large percentage of GRASS 5 commands now respond to the
"--interface-description" flag by outputting XML describing the command,
the options they take, etc.  This can be used to dynamically generate
interfaces.
  Some caveats:
    
    * Not all GRASS commands respond appropriately.
    * Some GRASS commands will respond appropriately, but still expect
      to take user input from the console.
    * Some GRASS commands will output different XML depending on the
      output of various function calls.  This is mostly used to collect
      a list of "options", and is most frequent with the display
      commands.

I've mucked with this a bit using XSL Transforms.  Have a look:
  http://www.jps.net/egm2/grass/xml

The XML thing is still a work in progress.  I'm playing with it right
now to generate an interface via the GTK toolkit (more a learning
experience in GTK than anything...).  There's an implementation of
dialogs via Python/wxWindows in the GRASS sources.  With some more work,
the XML thing could pan out nicely, but it does need a bit more work...

-- 
Eric G. Miller <egm2 at jps.net>