Helena wrote:
But I don't think anybody is actually trying to compute and visualize
TIN surfaces like this because the raster DEM from the same points
is just so much better.
they'll still try if it's what they've grown up with 
Going back to nviz - Google SoC has been announced, I lost
track where it stands now
your timing is very good, Wolf should announce something very soon.
I'll note that applications for mentoring organizations have not yet
opened and no one is guaranteed a spot or a specific number of students.
Still it helps to start thinking about this stuff early anyway.
but I am wondering whether we should set up 2010 topics wiki
done. http://grass.osgeo.org/wiki/GRASS_SoC_Ideas_2010
as you add ideas, keep in mind that we need both a great application
AND a matching mentor to accept anything- so in particular I worry about
who could mentor the much needed imagery & orthophoto port to wxPython.
Handing the student the "too hard" pile and walking away is rather risky.
also, for ideas with strategic and complex integration issues (C++ etc),
we need to specify how we want it to happen up front, rather than have
the student do all this hard work and then say sorry, it needs to be
rewritten to be merged into the main tree. I guess that's really a
project-wide communication issue, but it really helps in evaluating the
applications if we have an idea of what is viable and actually wanted
before hand. In practice focusing on "we want feature X" is secondary
to the quality of the applications we get; sometimes great students can
be convinced to put in multiple applications if we are impressed by
them but are not sure of their chosen task.
We (me and a group of students here) can certainly help to
test, develop manual/help pages
and maybe even do some programming to help get it moving to
replace nviz
which has many things broken or working only partially
(points with attributes, file sequencing tool,
isosurfaces, etc ).
what's the file sequencing tool do? is it like xganim in 3D but
simpler than the keyframe animator(s)?
(I wonder how hard it would be to quick-fix those things in the tcl?
It seems like a series of small bugfixes to me, just requires time and
tcl smarts to be thrown at it. and the discipline not to extend it 
Hamish