[GRASS-dev] Improving of GRASS GIS Errata proposal

[Anna Petrášová]

On Sat, Mar 26, 2016 at 7:25 AM, Maris Nartiss <maris.gis at gmail.com> wrote:
> Hello all devs,
> following recent movement into providing a greater transparency on
> GRASS GIS analysis trustworthiness, I have created a first draft of
> GRASS GIS errata framework proposal.
> https://trac.osgeo.org/grass/wiki/RFC/5_Errata
>
> Everyone is invited to improve it and provide feedback to prepare
> final proposal.
>
> At the moment there are two recent cases of GRASS GIS modules
> providing reasonably looking but incorrect analysis results. They both
> can be used as use-cases while evaluating need of such framework and
> its content:
> https://trac.osgeo.org/grass/ticket/2917
> https://trac.osgeo.org/grass/ticket/2969

I believe there is a third candidate:
https://trac.osgeo.org/grass/ticket/2820

v.surf.idw had 2 problems:

from the ticket:

Ok, found part of the problem: distance was always left as dy*dy +
dx*dx without ever taking the square root. This obviously changes the
weights as 1/10 is closer to 1/100 than 1/100 is to 1/10000...

The second is more difficult to describe, both fixed in r67211

Using the -n flag gave correct results, otherwise the results were
always wrong. The bug is present in GRASS 7.0.0 to 7.0.2.

Anna


>
>
> WBR,
> Māris.
> _______________________________________________
> grass-dev mailing list
> grass-dev at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/grass-dev