[GRASSGUI] qgis interface alike

Daniel Calvelo dca.gis at gmail.com
Sun Mar 25 15:11:31 EDT 2007


On 3/25/07, Michael Barton <michael.barton at asu.edu> wrote:
> The rationale for the plugin architecture in QGIS is that is it
> fundamentally a map display program, rather than a full-featured GIS. Using
> the plugins is a way to embed analytical and map development capabilities
> into QGIS while maintaining the basic architecture.

Yes. That helps enormously when trying to introduce somebody to it.

> However, to make this work, you need to have QGIS and GRASS installed.

That was my point: to leverage the GRASS plugin toolbox, you don't
need QGIS; all you need is the "definitions" for the plugin commands
and some programming. Those definitions are the equivalent of
menudata.py in current wx effort. They define an iconset, a subset of
grass-command parameters and a set of default values, so there is not
a one-to-one mapping between those plugin-commands and grass-commands.
It adds a layer of flexibility.

And it could be further expanded, in terms of usability. Consider e.g.
r.slope.aspect ; in the qgis plugin, in becomes two plugin-commands:
r.slope.aspect.slope and r.slope.aspect.aspect, which produce
respectively only the slope and only the aspect by feeding the
appropriate parameters to the grass-command. What do you gain? You can
default input to the select "layer", and you can default output to a
machine-generated name. So you do it all in one button click.

> The way wxGRASS (and the TclTk) architecture works is to wrap all of GRASS
> in a GUI, rather than build a GUI and then add in some GRASS. I'm not sure
> what would be served by building a GUI wrapper for GRASS that only accesses
> a limited part of the code base.

I was trying to discuss the reuse of Radim's "GRASS toolkit" idea
within our current effort, not suggesting reimplementing QGIS.

> What might make more sense, in terms of flexibility, is to make selected
> GRASS commands also available (i.e., beyond the complex menu hierarchy) via
> a graphical 'tool box', as Trevor Wiens has suggested. Since trying to
> create a tool box for all 300+ commands would result in an interface as
> complicated as the menus themselves, maybe we could make a tool box that
> lets users 'plug in' or remove different command groups (e.g., hydrology or
> terrain).

That's what I meant by "Can the qgis plugin interface make sense as a
"quick toolbox" alongside the heavier wxgui?". My concern is whether
and how that would confuse users (especially newcomers) instead of
giving them a gentler learning curve.

I like the idea of having task-related or thematic groupings. Another
fine idea from Radim is the building of the command icons with a very
simple symbology; that might help making those thematic toolboxes
friendlier.

Thanks,

Daniel.

> Michael
>
>
> On 3/25/07 9:59 AM, "Daniel Calvelo" <dca.gis at gmail.com> wrote:
>
> > Hi all,
> >
> > The current wx-based development of a GRASS GUI follows the line of
> > tcltkgrass, d.m, gis.m and provides graphical interactive access to
> > the whole GRASS functionality.
> >
> > The approach taken by Radim in the QGIS GRASS plugin is rather
> > different: he provides full CLI access for advanced use on the one
> > hand, and on the other he provides graphical shortcuts for common
> > operations, needing a reduced set of parameters (the rest using either
> > defaults or special values according to each "shortcut").
> >
> > I was thinking about leveraging Radim's work, building a Python parser
> > for his .qgm format, basically reusing qgis-grass plugin's module tree
> > and providing this same "fast" commands for the wx-based interface.
> >
> > But we're talking about different philosophies here. How do you think
> > both could be merged in a way that wouldn't be unintuitive to a new
> > GRASS user? Would separating the interface in "basic" and "advanced"
> > functionality be desirable? Can the qgis plugin interface make sense
> > as a  "quick toolbox" alongside the heavier wxgui? Would this
> > integration be worthwhile from a user standpoint?
> >
> > Daniel.
>
> __________________________________________
> Michael Barton, Professor of Anthropology
> School of Human Evolution & Social Change
> Center for Social Dynamics & Complexity
> Arizona State University
>
> phone: 480-965-6213
> fax: 480-965-7671
> www: http://www.public.asu.edu/~cmbarton
>
>
>


-- 
-- Daniel Calvelo Aros




More information about the grass-gui mailing list