[Rd] 3D visualization using vtk library

Peter Dalgaard BSA p.dalgaard@biostat.ku.dk
27 Mar 2002 12:19:36 +0100

"Jun Sung" <jun_sung@hotmail.com> writes:

> has there been any attempt to improve the 3D visualization (interactive)
> features in R, using the vtk library?
> It seems fairly straight forward.
> maybe even trivial for somebody who has experience with vtk library.
> let me know if there are any obvious reasons i may have overlooked

The same kind of thoughts have crossed my mind, but I don't think
anyone has actually tried it. It is certainly attractive, not just
because the pictures are pretty, but because it offers an interface at
a higher level than e.g. OpenGl (vtk is, for instance, object-based
rather than pixel-based).

It would be very useful if someone cared to sit down and create a few
simple examples.

Some things are clearly trivial since you can access a Tcl interpreter
directly from R via the tcltk package so whatever you can do in Tcl
you can do in R too. You can do it via .Tcl if you must, but you can
also create R commands that parallel the Tcl ones in the same way as
tkbutton() and friends. 

In fact, that interface is a pretty straight subroutine mapping and
actually not very Tcl-ish, so it would probably be rather easy to make
a mapping directly into R as it has already been done for Python and
Java. Easy in principle, that is. You would need an automated
interface generating tool because of the sheer size of the library.

What I don't quite see is how the data interfaces are pieced together.
If you have a data array in R, how do you get it across to the
renderer? This might be easy, but you need to dig deeper to find out.
The non-trivial objects in the example codes seem to be reading from
files, but surely they can also get the information from objects in
their hosting language. 

Much harder is the generic discussion about how you would want to
structure a set of 3D methods for statistical data, so that it becomes
simple and yet flexible in the same way as our current 2D methods (and
even those are under some attack because of the "indelible ink" aspect
and general lack of interactivity).

   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard@biostat.ku.dk)             FAX: (+45) 35327907
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch