[GRASS-dev] Re: gis.m zooming: 2 new bugs

[Michael Barton]

> 
>> This is getting close to finding out what it is you are wanting to see.
> 
> Michael,
> 
> It's not about me "wanting to see", but about fixing an issue that
> hampers gis.m. I wouldn't bother messing with nice folks just to
> enforce some silly feature everybody could live without.

It is often more helpful to find out how you think the interface "should"
behave than what you think is "wrong"--especially in this case where it is
necessary to write some (for me at least) conceptually complicated equations
to make zooming happen in different ways and under different conditions.

> 
>>> That's not what I mean. I mean that the geometry must be preserved
>>> exactly the way I set it by drawing the zoom box.
> 
>> So in the example you gave what is it that you think ought to be displayed.
>> I'm not trying to be a problem, but this is unclear.
>> 
>> You first used a zoom in box to set a region geometry that was tall (N-S)
>> and narrow (E-W). So far so good.
>> 
>> Then you draw a zoom out box that is smaller in the N-S dimension than the
>> current region and larger in the E-W dimension than the current region.
> 
> Correct.
> 
>> The algorithm increases the N-S region extents so that the parts of a
> map at the
>> previous north and south edges now fit into the box that you drew on the
>> screen.
>> 
>> What is it you think should happen with the E-W given the box you've drawn?
> 
> It should be extended accordingly, like N-S is, including the invisible
> part od the display (beyond the region, to the E). So that the
> resulting region reflected the proportions of the zoom box.

So you agree that the N-S region extent should be increased (well beyond the
box) so that the old region edges fit into the box.

In the E-W direction, you'd want the new region extents to match the box's
geographic extent rather than behave like the changes to the N-S extents.
This would only apply, however, only if the box is larger than the region
extents in a given dimension? in a given direction?


...snip snip...

>> 
>> Try the change to mapcanvas.tcl and see what you think.
> 
> I tried it and told you what I thought in my previous email. Have you
> read that?

It was not clear that you had looked at it yet.

> 
> Maciek

Keep in mind that with any other program we wouldn't even be having this
discussion. It's a function of the complex, sophisticated, and insanely
frustrating way that GRASS manages its region of interest.

Michael 


__________________________________________
Michael Barton, Professor of Anthropology
School of Human Evolution & Social Change
Center for Social Dynamics and Complexity
Arizona State University

phone: 480-965-6213
fax: 480-965-7671
www: http://www.public.asu.edu/~cmbarton