A successful research project requires having 3 components: Theory, Model and Data. The definitions of these words, relying on the conversations with others are these.
Theory is the concept or idea that leveraged or to be tested (demonstrated).
Model is the system or a series of processes such as an algorithm that implements or tests (demonstrates) the theory.
Data is the test results demonstrating the efficacy of using the model, thereby also confirming the theory.
In the paper on Shape-Carving by Kutulakos, the theory is the principle of having each ray of light from the surface of the object agree in color, and by starting with a volume larger than the actual volume of the object and carving away voxel by voxel, the outcome is necessarily an overestimate of the actual object volume. The model is the algorithm that implements the shape-carving algorithm. The data are the volumetric reconstructions resulting from this method as well as any special parameter value used in the tests. Since the data shows that the method can compute the volumetric reconstructions successfully, the data also confirms the theory that reconstructions can be obtained by carving away at a volume larger than the actual object by keeping only voxels that agree in color among its many camera projections.
Data doesn’t come as readily or as cheaply as the model and theory, apparently. For that reason, the data is what separates researchers at the cutting-edge and researchers striving to reach there.