Re-did the tutorial on finding transformation between 3D points in the note section. Fixed some mistakes in the maths and explanation. Amazing how often I get that simple algorithm wrong.
Got around to playing with VisualSFM. Need to tweak it to use less memory. It appears to give good results and is very fast. If this program works out for my need it will probably make RunSFM obselete!
Next thing on my TODO list is to do a follow up post on the matrix library comparison benchmark with any debugging/error checking turned off to see how fast it can go.