How Rust Makes Advanced Type Systems Accessible to the Masses
Talk Outline
These pictures are from 8-9 years ago!
Python
- Great for learning and teaching
- Online resources, tutorials, books
- English-like syntax with "and", "or", and "not" which is readable for people with little or no programming background
- Big ecosystem of libraries and many production users
import random
min = 1
max = 6
roll_again = "yes"
while roll_again == "yes":
print "Rolling the dices..."
print "The values are...."
print random.randint(min, max)
print random.randint(min, max)
roll_again = raw_input("Again? ")
if (roll_again.startswith("never") and
roll_again.endswith("ever")):
print "I didn't want to play anyway."
elif not (roll_again.startswith("no") or
roll_again.startswith("No")):
print "That wasn't an option!"
Python
Interpreted
print "Favourite Animal Test"
prompt = "What is your favourite animal?"
fav = raw_input(prompt)
response = "That is a "
if fav == "dog":
response += "great"
elif fav == "cat":
response += "cool"
elif fav == "dragon":
response += "unreal"
else:
response += "awesome"
response += " choice!"
print response 32 + "abc"Dynamically Typed
This isn't valid
The Python interpreter evaluates programs line by line
# This will produce some text
ans = raw_input("Pick some letters: ")
# This will produce an integer
ans2 = int(raw_input("Pick some numbers: "))
# You can't add text to an integer, but
# Python will still run the above code
# before failing on this next line
print ans + ans2# Dynamically typed also means that
# you can do something like this:
stuff = "abc"
stuff = 22
# stuff can be anything you want and
# the Python interpreter will only
# check it when you use itTypes
Types and type systems are a way to provide more information about your code to a program that can check your code automatically.
Program with Types
Type Checker
User
Feedback
Program
Types
You don't need types!
- A human can check pretty much everything that a type checker could check for you
- The question is: Why should we?
- It is easy for computers to check a lot of these details for us
- We just have to give them enough information to do that
- Why spend your time doing something you can automate away?
Dynamically Typed
- Python
- Ruby
- JavaScript
- Racket/Lisp
- Lua
- PHP
- and many more!
Statically Typed
- C, C++, C#
- Java
- Rust
- Haskell
- Scala
- Swift
- and many more!
Types are checked while the program is running
Types are checked before the program can run
Rust
- Rust is a statically typed, compiled language
- In development since 2006, stable 1.0 version released in 2015
- Most loved programming language in the StackOverflow developer survey two years in a row in 2016 and 2017
- Fantastic community full of kind, compassionate and very knowledgeable people
- Lots of great learning resources, thorough documentation, and people who are happy to answer your questions
Rust
- Review: Types and type systems are a way to provide more information that can be checked beforehand
- Rust checks for trivial errors like passing the wrong types to a function or using incompatible types with an operator
- Programs that type check are guaranteed to be memory safe and not have any data races
- You can write code that will not type check unless it is correct
Rust
- The speed of C and C++, memory safety of Java, and the expressiveness of Haskell
- Clean and relatively simple syntax that lets you focus on what you want to do and not need to write a lot of extra code to do it
- From the FAQ: "We do not employ any particularly cutting-edge technologies. Old, established techniques are better."
Friends of Rust
https://www.rust-lang.org/en-US/friends.html
Takeaways
- Rust makes advanced type systems accessible to everyone by having a great community, a fantastic set of features, and excellent learning resources
- Type systems are good, important parts of programming languages that will reduce the amount of problems in your code
- We should be using advanced type systems so that we can give the computers checking our code as much information as possible
- The Rust Book
- Rust by Example
- Rust Cookbook
- Programming Rust
- Rustlings
- Rust Learning
- into_rust()
- and many more resources that can be found using your favourite search engine!
Learn Rust
- Learn Rust from scratch by drawing pictures
- No prior programming experience required
- Ready in a few weeks!
Turtle
Website: turtle.rs
Thank you!
How Rust Makes Advanced Type Systems Accessible to the Masses - StarCon UW 2018
By Sunjay Varma
How Rust Makes Advanced Type Systems Accessible to the Masses - StarCon UW 2018
In the past it was accepted that only experts and highly experienced programmers could work on really large codebases. Those people all had the “special” capacity to keep all the different parts and considerations in their mind so that they could avoid a whole host of issues that a less experienced person might make. Nowadays, that convention isn’t sufficient. There’s too much software and too few “experts” to work on it all. Everyone has their limits and we need something more to make sure our software is correct. Why not let our computers do most of this checking for us? Type systems are a way to annotate our programs so that computers can do a lot of the work we used to rely on experts for. Advanced type systems have been around for a while, but they haven’t seen much adoption until recently. In my talk, I will cover this transition and talk about what makes Rust so different. From individuals to large organizations, people are all of a sudden jumping to adopt Rust everywhere. Rust has been the most loved programming language in the StackOverflow Developer Survey for two years straight. It’s type system allows any developer regardless of experience to modify a codebase with confidence. They know that they can lean on the compiler to tell them their mistakes. It’s not a perfect way to eliminate all errors, but this exciting transition towards types will give us a route towards software that is extremely maintainable and even fun to write for years to come.
