Choosing a Language

• You are an economics grad student and need to choose your weapon of choice.

Taxonomy of Languages

Hello, World?

C/C++

• The world pretty much runs on C++.

C++ Code Example

//sumvec.cpp
#include <iostream>
#include <vector>
int main(){
    std::vector<int> x;
    for (int i=1;i<5;i++){
        x.push_back(i);
    }
    int sum = 0;
    for (std::vector<int>::iterator i=x.begin();i!=x.end();i++){
        sum += *i;
    }
    std::cout << "sum is " << sum << std::endl;
}
//compile
g++ sumvec.cpp -o sum.x

C sum

• Defining the function:

  #include <stddef.h>
  #include <stdio.h>
  double c_sum(size_t n, double *X) {
      double s = 0.0;
      for (size_t i = 0; i < n; ++i) {
          s += X[i];
      }
      return s;
  }
  
  //main
  int main(void) {
      double x[] = {1, 2, 3, 4, 5};
      size_t n = sizeof(x) / sizeof(x[0]);
  
      double result = c_sum(n, x);
  
      printf("C-computed Sum = %.2f\n", result);
      return 0;
  }

C++: Pros and Cons

Python

• The general purpose language out there. Swiss Army Knife. All Terrain. 🚙

python sum

• we just use the built-in sum:

  a = [1,2,3,4,5]
  sum(a)

• that’s it!

Python Pros/Cons

Python Performance Analogy

R

• High level open source language for statistical computing.

R sum

• we just use the built-in sum:

  a = 1:5
  sum(a)

• that’s (again) it!

R Pros/Cons

Fortran

• The Grandfather of all languages 👴🏽

Fortran Example

• create a text file test.f90:

  program sumit
      implicit none
      double precision a(5)   ! allocate an array with 5 slots
      a = (/1,2,3,4,5/)       ! fill with values
      print *,"result is ", sum(a)
  end program sumit

• compile it and run it:

  $ gfortran test.f90 -o test
  $ ./test
   result is    15.000000000000000  

FORTRAN Pros/Cons

Victims of Speed vs Productivity Tradeoff?

• We have seen high-level langs (R, python etc) are good for productivity: Iterate fast on try, fail, repeat

Case Study

🔥🔥 Our Time is Running Out 🔥🔥

Oswald (QE 2019): Homeownership and Location Choice

When the Clock is Ticking ⏰

When the Clock is Ticking ⏰

Why We Created Julia - 2012 blogpost

We are power Matlab users. Some of us are Lisp hackers. Some are Pythonistas, others Rubyists, still others Perl hackers. There are those of us who used Mathematica before we could grow facial hair. There are those who still can’t grow facial hair. We’ve generated more R plots than any sane person should. C is our desert island programming language. We are greedy: we want more.
We want a language that’s open source, with a liberal license. We want the speed of C with the dynamism of Ruby. We want a language that’s homoiconic, with true macros like Lisp, but with obvious, familiar mathematical notation like Matlab. We want something as usable for general programming as Python, as easy for statistics as R, as natural for string processing as Perl, as powerful for linear algebra as Matlab, as good at gluing programs together as the shell. Something that is dirt simple to learn, yet keeps the most serious hackers happy. We want it interactive and we want it compiled.

Julia History of Success

• That proposal was very bold.
• Not the first time that a new language arrives. It’s hard to carve out some market share.
• Development started in 2009
• Version 1.0 released in 2018
• Funding model to preserve language development (JuliaComputing Inc)

What is Julia?

• Modern language

Benchmarks

Which problem does Julia want to solve?

• Julia cofounder Stefan Karpinski talks about the 2 languages problem

Which problem does Julia want to solve?

Scientific Software Stacks

Comparing Programming Languages in Economics

• S. Boragoan Aruoba and Jesus Fernandez-Villaverde compare performance of languages on an Industry Standard Growth Model.

Results

Not Mentioned

There are many other languages out there. We have not mentioned

• Matlab: widely used in economics, but clearly inferior to julia (Expensive + license block)
• Stata: very popular in microeconometrics, old design, commercial software.
• Java
• Rust
• Scala
• Go