Hi Victor,
A great idea, the original one :-). I am sure you are right the way data
is organized in GRASS is scary for new users. It took me some time
before I got used to GRASS and its way of data organization. And
although I like it very much now, it is a steep learning curve. So this
would really open the GRASS set of functions for a much wider audience.
I am also happy you are planning the other way around for the near
future. Unfortunately I cannot offer much help in terms of programming,
but when the time is there, I am happy to do some testing if needed.
Best regards,
Paulo
On 04/02/2012 12:10 PM, Victor Olaya wrote:
Paulo,
The idea (at least, my original idea) was to let users with no GRASS
knowledge to use GRASS algorithms from a more common GIS interface. So
they do not have to worry about mapsets and things like that, and they
execute the algorithm just like they would do with any other one. The
mapset and location is created automatically each time the algorithm
is run, and all data imported into it. Then the results are exported
and loaded in the GIS.
This might have some limitations, of course, but in the Java version
of SEXTANTE it has proved very useful to make those algorithms
available to most users, specially those that were a bit afraid of the
complexity of GRASS or were not used to its interface or its
command-line syntax.
Using GRASS layers is a different thing, but as I said, I hope to have
that already in a near future, and also to be able to use those GRASS
layers in other algorithms, so SEXTNATE will become the intermediate
layer for all analysis and data sources, and doing it as transparently
as possible
Regards
El día 2 de abril de 2012 11:54, Paulo van Breugel
<p.vanbreu...@gmail.com> escribió:
On 04/02/2012 09:47 AM, Paolo Cavallini wrote:
On 30/03/2012 23:19, Victor Olaya wrote:
Well, the idea is just the opposite: to use any raster or vector layer
in GRASS, directly from QGIS. "Any" in this context means any "normal"
Please remember also that it is possible to use rasters in GRASS without
importing them, but just registering them (r.external), with insignificant
loss of performances. Thus, we can more or less forget about grass rasters
in their native format, for most of the uses.
All the best.
Just to be clear, I assume you mean here that there is no need to convert
(import) other raster layers to GRASS layers in order to use GRASS
functions? I am asking because this option doesn't help the user who has
his/her data in a GRASS database.
For me the strict and transparent handling of region settings and projection
has always been one of the strong points of GRASS, and I am curious how the
workflow will be when working with external rasters. Like with internal
rasters, an external raster layer does need to be in the same projection as
the current location. Furthermore, raster functions use the region settings
(extend, resolution) of the current mapset as far as I understand.
When importing, it is possible to create a new location on the fly based on
the raster you are importing. This makes it easier to deal with the above (I
think), e.g., by creating a temporary location. I am not sure linking an
external raster offers the same option (I had a quick look, but can't find
the option, but maybe I just overlooked it). But then, the normal import
function allows to create a location based on a raster without actually
importing the raster. So the two could be combined.
_______________________________________________
Qgis-user mailing list
Qgis-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/qgis-user
_______________________________________________
Qgis-user mailing list
Qgis-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/qgis-user
