SLIDE 1
Haskell Deian Stefan (adopted from my & Edward Yangs CSE242 - - PowerPoint PPT Presentation
Haskell Deian Stefan (adopted from my & Edward Yangs CSE242 - - PowerPoint PPT Presentation
Haskell Deian Stefan (adopted from my & Edward Yangs CSE242 slides) Why Haskell? The great ideas [Haskell] Expressive power (say more with less) Pattern matching First-class functions Exception handling Type inference Continuations
SLIDE 2
SLIDE 3
The great ideas [Haskell]
Expressive power (say more with less) First-class functions Type inference Monads Pattern matching Exception handling Continuations Reliability and reuse Type polymorphism Modules Objects & inheritance Type classes Cross-cutting concerns Memory management Concurrency
SLIDE 4
What is Haskell?
a typed, lazy, purely functional language
SLIDE 5
Haskell is statically-typed
SLIDE 6
Haskell is statically-typed
- Everything has a type
- Everything must make sense at compile time
➤ Unlike JavaScript where f(x) with f=undefined will
not complain until you actually evaluate f(x)
- Is JavaScript typed?
➤ A: yes, B: no
SLIDE 7
Why is this cool?
SLIDE 8
Why is this cool?
- Removes whole classes of bugs
- Address bugs early vs. after they have been triggered
➤ Prevent weird errors from creeping up on you ➤ Important for safety, security, and compositionally
- Easier to optimize and write faster code
➤ You can remove your typeof checks; compiler can do
fast things. V8 relies on types to makes things fast!
SLIDE 9
Haskell is functional
- This means no“side-effects”?
➤ A: yes, B: no
SLIDE 10
Haskell is functional
SLIDE 11
Haskell is functional
- Support for high-order, first-class functions
- Meaning of programs centered around:
➤ evaluating expressions ➤ not executing instructions
SLIDE 12
Haskell is pure
SLIDE 13
Haskell is pure
- Expressions (e.g., functions) don’t have “side effects”
➤ Is JavaScript pure? A: yes, B: no
- Everything is immutable: mutation is a side-effect!
- What does it mean for an expression to not have side-
effects?
➤ In scope where x1, …, xn are defined all occurrences
- f e (where FV(e) = {x1, …, xn}) have the same value
SLIDE 14
Why is this cool?
Don’t take it from me, take it from Backus
SLIDE 15
Why is this cool?
- Algebraic laws: equational reasoning & optimizations
➤ Can always replace things that are equal, λ calculus!
- Easier to think about
➤ e.g., don’t need to worry if x changed after calling f
- Parallelism
➤ Can evaluate expressions in parallel!
SLIDE 16
Haskell is lazy
SLIDE 17
Haskell is lazy
- You don’t evaluate an expression until its result is
absolutely necessary: in contrast to JavaScript
➤ Remember: call-by-name
- Haskell’s evaluation strategy is called call-by-need
➤ Because of the other properties: you actually only
evaluate an expression once and cache the result
SLIDE 18
Why is this cool?
SLIDE 19
Why is this cool?
- Can define your own control structures using functions
➤ E.g., defining if-then-else is much easier in Haskell
can be done naturally; less so in JavaScript; why?
- Can define infinite data structures
➤ E.g., infinite lists, trees, etc. ➤ Can solve general problem and then project solution
SLIDE 20
Haskell is a committee language
SLIDE 21
Why is this interesting? [SPJ]
1yr 5yr 10yr 15yr 1,000,000 1 100 10,000
Geeks Practitioners
C , C + + , J a v a S c r i p t , R u b y , . . .
SLIDE 22
Why is this interesting? [SPJ]
1yr 5yr 10yr 15yr 1,000,000 1 100 10,000
Geeks Practitioners
Research languages
SLIDE 23
Why is this interesting? [SPJ]
1yr 5yr 10yr 15yr 1,000,000 1 100 10,000
Geeks Practitioners
Successful research languages
SLIDE 24
Why is this interesting? [SPJ]
1yr 5yr 10yr 15yr 1,000,000 1 100 10,000
Geeks Practitioners
Committee languages
SLIDE 25
Why is this interesting? [SPJ]
1yr 5yr 10yr 15yr 1,000,000 1 100 10,000
Geeks Practitioners
Haskell
SLIDE 26