Summary:
The paper describes and discusses the software written to control and test (w.r.t. accuracy) a custom-made image-guided robot system (presented at MICCAI 2006). The robot controller is based on a low-cost controller board. The software submitted together with the paper is based on several open source projects, and the part regarding the robot controller may be used as a practical example of how to use the controller board (Galil DMC-2143).

Open Science:
The source code is provided including the data acquired to produce the results presented in the MICCAI paper.

Reproducibility:
I downloaded the code, but did not try to compile it, because it is based on quite a lot of other packages, which I do not have installed on my computer (and I do not have a robot). It is clearly described in a readme file which tool requires which additional package.

Use of Open Source Software:
A lot of open source packages are used and (some of them) discussed: cisst, vtk, swig, python, wxPython, cmake, 3D Slicer, cvs.

Open Source Contributions:
The source code is provided (except the vtkRodent class integrated into 3D Slicer) and seems to be in a usable form (but I did not actually try to compile or use it). It is clearly described how to use the code (partially in the paper, additional details within the code or in readme files).

Code Quality:
The code is reasonably easy to read, documented, and in a modern coding style. It has been compiled and tested in the Microsoft Windows environment. It is not clear whether all parts may work on other platforms; the used external packages do work on other platforms.

Applicability to other problems:
Although the control software is specific for the described robot, it may be useful as an example of how to use the controller board. The accuracy calculation tools may be useful for similar experiments.

Suggestions for future work:
The planned integration of the part regarding the controller board into the cisstInteractive library is probably desirable.
FYI, regarding the region grower: vtk has also some image processing capabilities, e.g. for (simple) region growing the filter vtkImageSeedConnectivity.

Requests for additional information from authors:
I would be interested in a few more details on the experiment the robot was designed for (oxygen measurement in a 3D grid pattern). This is not really important, because its not the focus of the paper and the robot can be used for other purposes. Probably I have to read the MICCAI paper.