
* In the Mesh->Resolutions->Other dialog box showing the different
  mesh resolutions, selecting "Delete" sometimes results in an
  unexpected error message (especially when deleting the lowest
  resolution mesh).

* CyberScan Preview is meaningless when no CyberScans are loaded

* Can the IRIX version be made to compile with gcc/g++, instead
  of requiring CC/cc (which are expensive!) ???

* Why not just package things up into a single executable file,
  using FreeWrap, ProWrap, etc., for the binary distributions, to
  avoid all the hassle of the Tcl/Tk dependencies?

* Saving/loading image files doesn't work on Linux version
  (no IFL library)

* "Open SD scan directory" doesn't work properly on filenames
  such as "bogus..sd".  This might just be a TCL bug or version
  problem...the version of TCL linked with scanalyze evaluates
  [file extension bogus..sd] as "..sd", whereas when I just run
  tclsh, this expression evaluates to ".sd".

* If the Windows scanalyze.opt32.exe version is compiled using
  "make opt32", then the program will crash when asked to do
  a line depth plot.  This problem can be avoided by compiling
  the file "plvAnalyze.cc" separately, without the optimization
  flags.

* ZBuffer readbacks don't seem to work at all sometimes; this
  has been observed on both Windows and Linux when using GeForce3/4
  graphics.  It might be a graphics timing issue; often just rerunning
  Scanalyze a second time avoids the problem.  Printing out the
  min/max Zbuffer values that Scanalyze computes during initialization
  may shed some light on this problem.

* In plvMain.cc, see "BUGBUG" for forward declaration of isPlyFile. This is a
  compiler bug, it shouldn't be needed since it is in ply++.h

* In plvCmds.cc, see "BUGBUG" for forward declaration of class Scene. This is
  a compiler bug, it shouldn't be needed since it is in plvScene.h

