[GRASS-dev] d.m/d.gis mysteries..

[Joel Pitt]

On 10/31/06, Dylan Beaudette <dylan.beaudette at gmail.com> wrote:
> Moritz : Although I have not contributed to this discussion until now, this
> bug is something that bothers me as well. In fact, the implications of the
> behavior of d.zoom were not fully clear to me until Maciej pointed them out
> on the list one day. Since most of my work deals with point and raster data,
> strict adherence to the region settings are of the utmost importance. I now
> always check, and sometimes reset, the region with the -a flag to maintain
> meaningful (for me) settings.

I don't use gis.m, mainly because all my scripts interact with x
monitors. d.m is better suited as it is designed to work with the x
monitors - eventually I'll update my scripts though.

I know that if I used gis.m, and it responded in the way described in
this thread, I would be strongly discouraged from continueing using
it.

> Michael: I understand how difficult it can be to implement something that you
> need, especially when you are not a programmer by trade. I am in the same
> boat, and have found myself getting into trouble now and then when I dabble
> into developing my own tools. You have done a tremendous effort on the gis.m
> interface and should be proud. However, Maciej is correct in pointing out
> that there is a bug, and that it needs to be addressed sooner than later.
> Nothing puts people off more than inconsistent, or even worse, inaccurate
> behavior of some product. Not to say that gis.m isn't ready for general use,
> rather this issue really should be tackled with the help of the core
> development team.

I don't have time myself to investigate the bug, and it'd take me a
while to understand it since I don't know TCL. However, from what I
understand it it seems like you could present the user with the true
extent values from the zoom, but when drawing the PNG you could round
up the extents when zooming, and maybe add the region resolution to
each extent. Changing the region is a pretty quick process so doesn't
matter if there are multiple calls per zoom.

I guess the problem with this would be parts of vector objects that
are not in the true region showing up in png. i.e. a large vector
point icon might show up due to the icon size overlapping into the
true region.

Just some ideas - possibly you've already thought of this and I'm
missing something important...

cheers

P.S I don't mean to be purile, but surely "gis.m" isn't the best name
for the display manager when the g could be pronounced with a j sound.
Maybe [gd].mgr, [gd].gis, or [gd].grass?

-- 
-Joel

"Wish not to seem, but to be, the best."
                -- Aeschylus