[GRASS5] A naive opinion on how grass *should* work

[Eric G. Miller]

On Fri, May 03, 2002 at 10:11:54AM -0400, Russell Nelson wrote:
> Glynn Clements writes:
>  > Even more significantly, it's also used when reading existing raster
>  > maps. With a few exceptions, input maps are resampled according to the
>  > current region settings.
> 
> Didn't people just tell me that I could change it any time I wanted?
> How can I do that if my input maps got resampled?

To clear up confusion: pretty much all import routines use the extents
and resolution of the source data.  I'm not sure if there are any that
don't behave that way.  And, pretty much every raster operation uses
the current region except where the semantics dictate working on the
whole extent/resolution of the raster (r.null, etc...).

> GRASS should *always* do the right thing by default, and say what it
> did.  If a user asks it to do something that doesn't make sense, GRASS 
> should do it anyway, and print a warning.

Just having this discussion proves there isn't necessarily a "right
thing."  You may have certain expectations based on other experiences
but other folks are just as likely to have different expectations.

>  > But *most* programs need a region. The region concept is so pervasive
>  > that it just isn't worth complicating the majority of programs to
>  > allow for the possibility of there not being a current region. 
> 
> You could be quite right.  Please respect my newbie experience telling
> you that region-setting is difficult and confusing.  The defaults are
> not reasonable and have no relationship to the data I'm trying to
> import or view.

It is a recognized fact, that the region concept trips up people new to
GRASS.  However, it doesn't take long to get used to it, and it comes in
real handy at times as well (you'll just have to take my word on that ;)
 
>  > Especially when all that it would achieve would be to allow newbies to
>  > defer learning about the region for a minute or two.
> 
> Even if all they can do is run grass and type "help", that is a
> valuable addition.  I think you've forgotten how frustrating it is not
> even to be able to start up a program without needing to be an expert!
> I appreciate your expertise, and thank you for responding to my
> concerns.  But understand that my ignorance is valuable, too (while it
> lasts :).

Yes, GRASS could use some more introductory material.  There used to be
a command called g.help, but it got woefully out of date and nobody
wanted to work on it.  There've been some other discussions about
writing some tutorial documents that are current, so hopefully those
folks are charging ahead.

>  > It might be worth making it easier to automatically create a suitable
>  > location when importing a raster. But the main problem there is that
>  > the data being imported is frequently uncorrelated, so the projection
>  > and the coordinates *must* be specified by the user.
> 
> Unless it's already georeferenced.  I thought that was the whole point 
> of GeoTIFF -- to carry the projection and coordinates around with the
> raster data itself.

r.in.gdal fully supports GeoTIFF (among others).  There are import
routines for some image types that don't have such information. r.in.bin
supports generic raw binary formats, for instance.

-- 
Eric G. Miller <egm2 at jps.net>