gvx

[mbp]

> I find it easier to have gv on my path and execute it while I am in 
> the directory that has the file I'm trying to look at.

Of course.  I think the best we can do with geomview right out of the
box, assuming the user specifies no information, is arrange for it to
work from its own directory.  If you want to install it elsewhere you
can edit the "geomview" shell script; only one line has to be changed
and there are comments at the top of the file telling you what to do.
After changing the relevant line in "geomview" to indicate the
absolute pathname of the geomview directory, you can run it from
anywhere.  The new README file, which I finished updating after you
got gv0.tar.Z, explains this.  The README you have is the old one.


> There is one really annoying feature in the istropic geometry switch in 
> gv1.3, and that is when you switch geometry the particular view of the 
> world that took 20 minutes to get just right, is destroyed and the default
> values are reinserted. 
> 
> It would be much nicer to leave resetting to the reset button and just
> switch the geometry, like when you change the lighting, it doesn't change
> the window, the position, etc.

The problem is that quite freqently the view settings don't make sense
in the space you switch to.  The matrices describing the positions of
the geoms and cameras must be isometries (or similarities in Euclidean
space); switching spaces would break this unless they're all reset to
the identity, or to some space-specific default values.  To get
reasonable behavior close to what you want I think we'd have to store
a separate matrix for each space for each geom and camera, which
seems a bit excessive.

Plus, more to the point, why are you switching back and forth between
spaces with one data set anyway?  Does that action make mathematical
sense?

> The `inapplicable-button-syndrome continues to annoy, but after you 
> explained it to me I am aware that there's no help but to get used to
> it. 

I'm not sure what you mean by this.

> More problematic is the fact that there does not seem to be a way of
> seeing the entire data set in one view (however non-euclidean,
> non-hyperbolic, and non-elliptic that might be). I think that once I
> figure out whether such a view is even possible with gv, I will have
> to learn how to write .geomviews so that when a display list is
> first installed, it comes on in a total way (afterall, all the
> display lists are finite.).

Try selecting "world" in the object browser and then click on the new
"Look" button in the tools panel.  It arranges for the camera to look
at the selected object.

--Mark