Hi all,
In r.damflood if you try to run a calculation using invalid input you get a warning WRT Courant-Friedrich-Lewy stability condition. If the wikipedia article is accurate and it is indeed a necessary condition for convergence wouldn’t it make sense to raise a fatal error (or whatever is more appropriate) as soon as the condition isn’t respected? I mean I don’t see much point in running a potentially lengthy simulation that will give meaningless results. Am I wrong in this?
If my suggestion has indeed merit, could anyone suggest the most appropriate way to do this?I don’t mind submitting a patch but I am afraid that I am not really familiar with GRASS C APIs.
These are the relevant lines:
https://trac.osgeo.org/grass/browser/grass-addons/grass7/raster/r.damflood/SWE.c#L504
https://trac.osgeo.org/grass/browser/grass-addons/grass7/raster/r.damflood/SWE.c#L708
thank you,
Panos
Hi,
I’m agree with Anna. In my opinion if the input are correct the warning is only a warning and it is correct!
It is important in order to inform the user but the code is optimized in order to automatically change the computational timestep (lower or higher) in order to obtain the best performance without compromize the stability.
Perhaps you can do soem review n the html manual page in order to help users to insert correct input?
All the best and happy new year 
R
--
Eng. Roberto Marzocchi, PhD
R&D coordinator
Gter srl Innovazione in Geomatica, Gnss e Gis
Spin-off Progetto Uni.T.I. promoted by UNIGE
Via Greto di Cornigliano 6r - 16152 Genova
P.IVA/CF 01998770992
ph: 010-8694830 Fax: 010-8694737
mob: 349-8786575
E-mail: [roberto.marzocchi@gter.it](mailto:roberto.marzocchi@gter.it)
skype: roberto.marzocchi84
[www.gter.it](http://www.gter.it)
--
Gter social
[www.twitter.com/Gteronline](http://www.twitter.com/Gteronline) - [www.facebook.com/Gteronline](http://www.facebook.com/Gteronline) - [https://plus.google.com/+GterIt/posts](https://plus.google.com/+GterIt/posts)
[www.linkedin.com/company/gter-srl-innovazione-in-geomatica-gnss-e-gis](http://www.linkedin.com/company/gter-srl-innovazione-in-geomatica-gnss-e-gis)
-----------------------------------------------------------------
Please consider the environment before printing this email!
···
2015-12-31 1:16 GMT+01:00 Anna Petrášová <kratochanna@gmail.com>:
Hi,
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user
On Wed, Dec 30, 2015 at 6:40 PM, Panagiotis Mavrogiorgos <pmav99@gmail.com> wrote:
Hi all,
In r.damflood if you try to run a calculation using invalid input you get a warning WRT Courant-Friedrich-Lewy stability condition. If the wikipedia article is accurate and it is indeed a necessary condition for convergence wouldn’t it make sense to raise a fatal error (or whatever is more appropriate) as soon as the condition isn’t respected? I mean I don’t see much point in running a potentially lengthy simulation that will give meaningless results. Am I wrong in this?
I was getting this warning several times in the beginning of the simulation, but the simulation still produced reasonable results. Lowering the time step would probably eliminate it. I think the warning is appropriate and shouldn’t be changed to fatal error in my opinion, unless the original authors or anyone else who fully understands it say otherwise.
Best,
Anna
If my suggestion has indeed merit, could anyone suggest the most appropriate way to do this?I don’t mind submitting a patch but I am afraid that I am not really familiar with GRASS C APIs.
These are the relevant lines:
https://trac.osgeo.org/grass/browser/grass-addons/grass7/raster/r.damflood/SWE.c#L504
https://trac.osgeo.org/grass/browser/grass-addons/grass7/raster/r.damflood/SWE.c#L708
thank you,
Panos
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user