[GRASS-user] Scientific Comparison between GRASS and Proprietary Remote Sensing/GIS Software

[Hamish]

Firman Hadi wrote:
> One week ago I had a discussion with someone who is expert in remote 
> sensing.
> It was a really nice discussion though we have a different perspective
> in using software.
> I prefer to use GRASS, because it's an open source software while he 
> prefer to use proprietary software.


> He believed that GRASS can't do some tasks that proprietary softwares
> do.

It is impossible to match every feature when the opponents are a single
piece of software vs. an entire software business model.

GRASS can't do some things that proprietary GIS do, sure; but GRASS will
do more things than the vast majority of other GISs will (proprietary or
not). GRASS 6.3 is up to 390 modules, plus there are approx 50-100 addon
modules out there. It does a lot of tasks over many fields of endevour.

corollary- what are the licensing costs to buy every proprietary toolbox
needed to match the features of a full GRASS install?
Can it be done for < US$10,000? <$50,000? <$100,000??


> He thought that GRASS is not mature enough to be used.

Parts of the core libraries date back to the early 1980s. core
projection and import/export libraries are best of breed in the industry
AFAICT. At least for GRASS's raster map library (libgis), bug reports
are very rare. For those parts, you will not find much more mature
specialized software anywhere. Could it be more mature if it was 50
years old instead of 25?
(granted maturity ~ integral of time*userbase, not time alone. And maturity
is not the same thing as refinement+polish [prettiness])


> He asked me some prove that the results from GRASS is the same with
> proprietary software.

In the general sense, it is an impossible task to do this with any two
software. More important would be to test that the results are correct
for both softwares. This is best tested using in-house test cases which
deal with tasks which your organization works on; it is impossible prove
that GRASS is a 100% clone when it isn't trying to be. In many cases
GRASS will produce a more correct result, but different. In other cases
not, but to know you have to test. Broad assumptions and claims are
useless here.


> For me, it will be better if I bring actual comparison between GRASS
> and proprietary software, though maybe it will take some time. Is
> there any link or arguments to give an accurate explanation for such 
> kind of people?

you might try on the wiki site:
 http://grass.gdf-hannover.de/wiki/GRASS_Help#Migration_from_other_GIS_Software

In many cases I expect his proprietary software would not live up to the
standard he asks to be proven of GRASS.


good luck,
Hamish