I really enjoyed the point made on the first paragraph, especially the following phrase: > Some like SVMs for the elegance of their formulation or the quality

I really enjoyed the point made on the first paragraph, especially the following phrase: > Some like SVMs for the elegance of their formulation or the quality of the available implementations, some like decision rules for their simplicity and interpretability, and some are crazy about neural networks for their flexibility. Nowadays it seems that ANN (and related algorithms) get all the credit. It's refreshing to see someone point out why some people actually prefer other methods.