"The fact that we in computing keep trying to visualize introductory computing concepts is another piece of evidence for me that spatial ability is linked to su

"The fact that we in computing keep trying to visualize introductory computing concepts is another piece of evidence for me that spatial ability is linked to success with computing. We naturally try to use visualizations of algorithms and data structures in order to communicate how we imagine these computing processes and structures. However, there's some evidence that this is the wrong answer. A visualization may not be designed so that it's cognitively easy to process or easy to learn from. Worse, providing the visualization may prevent the student from internalizing the spatial mental model. You can only succeed in computing (my hypothesis suggests) if you can mentally transform your spatial models. If it's concretely in front of you, you may not develop that ability. Some of the SILC results suggest (to me, at least) that gesture may be a better way to communicate the spatial model. Gesture is ephemeral, and still serves to communicate spatial meaning." My take: in computing, "visualizing" how systems work actually requires thinking in N dimensions. If you try to represent something composed of a large number of dimensions in 2D or 3D, then of course you lose a lot of semantic context. But if you have to explain such an N-dimensional concept or structure to someone, of course you'll probably try to dump the model on a piece of 2D paper and the person you're explaining to will never really get it if he sticks to the 2D representation. "Internalizing" the model is simply reverse-engineering the 2D model back to the real N-dimensional model. If you try to represent something N-dimensional in M dimensions where M < N, you lose information and your new model will never be able to represent the full reality of the original model.