Critical Evaluation

We started working on the project at home and in the GrUVi lab. When we transferred our work to the CSIL computers, we were confronted with a series of segmentation faults that were not present elsewhere.

Fortunately, we used CVS, and were able to roll back our work and restart from a working state. Some or our segmentation faults came as a result of us not initializing some pointer attributes to 0, although oddly the program ran fine at home and at the GrUVi lab.
As mentioned before, time should have been allocated for communication between us. For the magnitude of this project, to ask someone to add a feature to it without learning the structure of the code well is two tasks in itself. I decided to not implement accurate color calculations so that I could have more time to work with Ian to set up certain rendering feature (i.e. curvilinear data rendering, and render by slice or region). In the end, I am glad I did so, because together as I team, we were able to accomplish more than if we each went our own ways.

Problems still exist in the program. As mentioned before, color calculations are not accurate as of yet. When I triangulated my tetrahedra, I calculated the two points on each tetrahedron between which maximum thickness occur. As a crude approximation, the color of the nearer point is calculated as an interpolated value between the vertices of the tetrahedron. That point is used for triangulation. Further, the vertices of each tetrahedron is not assigned an opacity of 0, but instead the opacity associated with the intensity at that vertex. Although this approximation is very crude, for the data sets that we rendered, recognizable images were produced. What we have now, along with the demos, have set up a great testing environment for the implementation of the color calculations later.
Another problem exist in triangulating the tetrahedra in curvilinear data sets. The calculation of the nearest point is wrong for certain viewing angles. The error occurs in either case 1 or case 2 tetrahedra (please see diagrams below), as suggested by debug messages. Further investigation is needed before we can determine where the problem originates.