3 ms·

Did you create a VM that is several times faster than Python's, for a static typed language thats as productive as a dynamic language, with true multithreading?

Did you create a VM that is several times faster than Python's, for a static typed language thats as productive as a dynamic language, with true multithreading? How did you outdo entire language communities?

imtringued 8y ago
That's a very low bar. It's just that Python programmers don't care about any of this. The productivity is already enough for them.
tomp 8y ago
I'm a huge proponent of typed languages, but there's no way that a statically typed language could be as productive as Python (with current type systems technology). For example, just the Pandas library offers incredible flexibility (often dynamically based on the arguments that different functions get) that I've yet to see replicated in a static language.
- brianpgordon 8y ago
  My experience is that you start out very productive not being weighed down by having to satisfy the type checker... but programs tend to grow over time and you end up regretting it. When you have a medium to large codebase in a dynamically typed language it gets to be more of a hindrance than a help.
  - tomp 8y ago
    I agree, but my point was rather that you can't typecheck Pandas (or a similarly complex library). Gradual typing for the win!
Aardappel 8y ago
Python's default implementation is so incredibly slow, such that any programmer that implements a new language that doesn't just make "everything dynamic" will automatically end up with a language that is faster than Python without really trying. There are probably a 1000+ language implementations out there that are faster than Python, most of which are single person efforts. Lobster is a much more static language, and does inlining of functions, inlining of structs, specialization etc that allow it to remove a lot of runtime cost, such that even the VM is fast. It also has been in development for about 8 years now. I've been working on compilers in one way or another for almost 30 years.

imtringued 8y ago

That's a very low bar. It's just that Python programmers don't care about any of this. The productivity is already enough for them.

tomp 8y ago

I'm a huge proponent of typed languages, but there's no way that a statically typed language could be as productive as Python (with current type systems technology). For example, just the Pandas library offers incredible flexibility (often dynamically based on the arguments that different functions get) that I've yet to see replicated in a static language.

brianpgordon 8y ago
My experience is that you start out very productive not being weighed down by having to satisfy the type checker... but programs tend to grow over time and you end up regretting it. When you have a medium to large codebase in a dynamically typed language it gets to be more of a hindrance than a help.
- tomp 8y ago
  I agree, but my point was rather that you can't typecheck Pandas (or a similarly complex library). Gradual typing for the win!

Aardappel 8y ago

Python's default implementation is so incredibly slow, such that any programmer that implements a new language that doesn't just make "everything dynamic" will automatically end up with a language that is faster than Python without really trying. There are probably a 1000+ language implementations out there that are faster than Python, most of which are single person efforts. Lobster is a much more static language, and does inlining of functions, inlining of structs, specialization etc that allow it to remove a lot of runtime cost, such that even the VM is fast. It also has been in development for about 8 years now. I've been working on compilers in one way or another for almost 30 years.