For me the biggest draws are a syntax that is closer to numerics and the C++ API for extensions. Scipy code looks clunky to me with a lot of distracting and irr

For me the biggest draws are a syntax that is closer to numerics and the C++ API for extensions. Scipy code looks clunky to me with a lot of distracting and irrelevant characters, but I'm probably just not used to it. Another difference is that Octave uses pass-by-value semantics with a copy-on-write implementation whereas I think scipy uses pass-by-reference. There are pros and cons to both approaches. I think in interactive use Octave is better able to recover from errors without trashing your data which is nice for quick calculations.