Numba vs. Cython: Take 2: The author compares different python implementations for N dimension pairwise distance. Numba stands out brilliantly: a single function decorator results in a 1300x speedup of simple Python code. I'm becoming more and more convinced that Numba is the future of fast scientific computing in Python. Two things makes it so fast: 1) no dynamic type checking; 2) a new
autojit that leverages the llvm; (Via: Nick Philips)
Why sequence the tuatara genome?: New Zealand scientists are starting to sequencing tuarara genome. Here is why it's so important: Modern birds descend from one branch in the diverse group we call dinosaurs, but each of those ten thousand species are dinosaurs. The tuatara, on the other hand, are the only living members of a lineage that separated from other reptiles more than 200 million years ago. (If you can't see the graph, it's here)
To justify having a computer on your face for everyone to see would have to deliver spectacular results.: if Google Glass would let you see 10x better than 20:20 vision then it might be a compelling proposition. The problem Google Glass has is that it simply isn’t much better than a smart phone, and it comes as a huge price – having something on your face.