Hacker News new | ask | show | jobs
by gypsyharlot 1969 days ago
I spent a few decades chasing "the next great language", and I regret it. I wish I had spent those hours learning about statistics or cryptography or algorightms. Things that are timeless. If you want to be a jack of all trades, master of none, by all means, chase the next great thing.

I for one, think it's more likely that Python will evolve. The mathematicians and physicists, biologists, astronomers etc. that have finally gotten rid of C++/Fortran and learnt Python aren't going to change to anything with forced static typing any time soon...
1 comments
nmfisher 1968 days ago
Not a mathematician/physicist/etc, but having literally dived into Julia today, I think you’d be surprised. It definitely seems like “Python 2.0” for numerical computing.
link
gypsyharlot 1968 days ago
Julia is probably excellent. How large are the benefits to using Julia compared to Python? Are the benefits great enough to compensate for the time spent:

- Training your entire team to use Julia instead of learning about other relevant things

- Re-writing your Python-infrastructure (or dealing with Julia to Python-interoperability, which means new employees now have to learn or know both Python and Julia)

- Replacing Jupyter and learning new, similar tools

- All employees making their favorite IDEs function well with Julia

- Figuring out how to make Julia talk to software that already has an existing Python API (Wireshark, AutoCAD, GNU Radio, ... just about everything you can think of)

link
nmfisher 1968 days ago
Those are definitely problems for the corporate world (except for Jupyter, a Julia kernel already exists for that).

Again, I’m not an academic so take what I say with a grain of salt - but I can see Julia paying dividends very very quickly in that sphere (once you’re over the admittedly steep learning curve).

link
kirill5pol 1968 days ago
A few of these are non-problems IMO.

- Julia is relatively easy to learn for Python people (especially the numerical people who have experience with stuff like Cython)

- The interop (both ways) is pretty great with Python

link
N1H1L 1968 days ago
The Ju of Jupyter is for Julia.
link