[GRASS5] setting the initial region

[Russell Nelson]

Glynn Clements writes:
 > Russell Nelson wrote:
 > 
 > > I believe that GRASS is made more difficult by the requirement to set
 > > the working region at start-up time.  I propose that the initial
 > > region should be "none".  Every program that uses the region should
 > > see that there is no region, and generate one that covers all the
 > > available data.
 > 
 > The term "every program that uses the region" is almost equivalent to
 > "every program". You are talking about re-writing most of GRASS.

Okay.  That doesn't scare me.

 > Programs which don't read any maps (e.g. r.mapcalc where the RHS
 > doesn't refer to any maps) cannot determine defaults for any of the
 > region parameters.

Then it would say "Please specify a region or set the default region
using g.region" and exit.

 > Programs which don't read any raster maps (e.g. v.to.rast) cannot
 > determine defaults for the resolution.

Then it would pick a reasonable one and report that it had done so.
Or if there's no reasonable way to choose one, it would say "Please
set the resolution of the region using g.region" and exit.

 > Programs which do read maps may still not be able to determine
 > defaults, as there may not be an obvious correlation between the
 > boundary and resolution of any input maps and suitable settings for
 > the output maps; e.g. r.proj.

There are probably more examples besides these.  They can be dealt
with by causing them to emit useful messages.

 > We already had this discussion a while ago, and the conclusion was
 > that your proposal wasn't feasible. Nothing has changed since then.

GIS has become more important, but GRASS has not become any easier to
use.  This is a problem, and it needs to be fixed.

-- 
--My blog is at angry-economist.russnelson.com  | Don't like poverty?
Crynwr sells support for free software  | PGPok | 
521 Pleasant Valley Rd. | +1 315-323-1241 cell  | Try economic freedom!
Potsdam, NY 13676-3213  | +1 212-202-2318 VOIP  |