[GRASSLIST:1506] Re: d.mon usage

[Eric G. Miller]

On Sat, Feb 24, 2001 at 08:46:30AM -0800, Rich Shepard wrote:
> On Fri, 23 Feb 2001, Eric G. Miller wrote:
> 
> > Almost never...  But if GRASS had a different architecture, I could see
> > the utility...
> 
>   Go on, Eric. What do you mean?

If grass had something closer to a "normal" GUI, I could see the utility
of having various views of data up simultaneously.  Like, maybe one view
has a broad area view, while another has a zoom window.  Or possibly
you'll want to look at a surface rendering side-by-side with a plan
view, with possibly the plan view having a box displaying the currently
shown view in the surface rendering.

I argued a while back that we should consider ditching the whole
monitor/module separation for display since it was designed to overcome
the widely varying display architectures way back when.  Since X is
pretty much the standard for unix boxen, we could move to having an
interface with a much higher level of functionality.  Actually, upon a
little research, I think Tk could be used for some windowing system
independence (NVIZ already does, though isn't the best UI).  Anyway,
I think a lot could be done in this area.  The main argument against was
the ability to script output to things like the D_cell driver.  IMHO, if
we also spent some time improving the ps.map system, then we'd still
have scriptable "image" output.  There are plenty of utilities for
converting PostScript to other formats (PDF, PNG, etc...).  The
postscript rendering model should be able to produce much better output
than something like D_cell could ever do.  Tk also has hooks for
producing PostScript, so the infrastructure is there...  I'm blown away
how good a little program like Sketch (Python and Tkinter) is, for
instance (and it's only at like version 0.6).

That's enough rambling...
-- 
Eric G. Miller <egm2 at jps.net>