‎Functional Programming on Perl‎

Christian Jaeger

LeafPair.com

Contents

Why functional programming?
Using the 'Functional Perl' project
'Functional Perl' project: plans
Perl 6, other languages?
Experimenting on the workstations

Contents

Why functional programming?
- slides--dense, show concretely below, ask then!
Using the 'Functional Perl' project
- demo--give intro
- show
'Functional Perl' project: plans
Perl 6, other languages?
- not well prepared :(
Experimenting on the workstations

Why Functional Programming?

Why Functional Programming?

"Difficult question.."

Why Functional Programming?

"Difficult question.."

Describing what it is, is not:

Why Functional Programming?

"Difficult question.."

Describing what it is, is not:

To program with functions that always return the same result for the same arguments (= pure functions).

fun square($x) { $x * $x }

Why Functional Programming? (2)

To program with functions that always return the same result for the same arguments (= pure functions).

Which implies not to use side channels

like less reasonable...

our $x;
sub set_x { $x= $_[0] }  # has a side effect
sub square { $x * $x }   # exposed to the side effect

and more reasonable ones.

our $count=0;
sub increment { $count++ }
sub count { $count }

# without side channels:
fun increment($count) { $count + 1 }
fun count($count) { $count }

Why Functional Programming? (3)

So, the question is, what is the motivation for such a style?

"Difficult to answer."

Why Functional Programming? (3)

So, the question is, what is the motivation for such a style?

"Difficult to answer."

"It makes making proofs easier"

Why Functional Programming? (3)

So, the question is, what is the motivation for such a style?

"Difficult to answer."

"It makes making proofs easier"

who the hell writes proofs?
and why does it make them easier?

Why Functional Programming? (3)

So, the question is, what is the motivation for such a style?

"Difficult to answer."

"It makes making proofs easier"

who the hell writes proofs?
- some type systems do them
and why does it make them easier?

Why Functional Programming? (3)

So, the question is, what is the motivation for such a style?

"Difficult to answer."

"It makes making proofs easier"

who the hell writes proofs?
- some type systems do them
- some brains do, implicitely
and why does it make them easier?

Why Functional Programming? (3)

So, the question is, what is the motivation for such a style?

"Difficult to answer."

"It makes making proofs easier"

who the hell writes proofs?
- some type systems do them
- some brains do, implicitely
and why does it make them easier?
- hm...

Goto considered harmful

Edsger Dijkstra, 1968 letter 'Go To Statement Considered Harmful' (originally 'A Case Against the Goto Statement')


sub Gsumofsquares {
    my ($from, $to)=@_;
    my $tot=0;
    my $i=$from;
  test:
    goto calculate if $i <= $to;
    return $tot;
  calculate:
    $tot+= $i * $i;
    $i++;
    goto test;
}


sub sumofsquares {
    my ($from, $to)=@_;
    my $tot=0;
    for (my $i= $from; $i<=$to; $i++) {
	$tot+= $i * $i;
    }
    $tot
}

Goto considered harmful (2)

The unbridled use of the go to statement has as an immediate consequence that it becomes terribly hard to find a meaningful set of coordinates in which to describe the process progress. ... The go to statement as it stands is just too primitive, it is too much an invitation to make a mess of one's program.

Goto considered harmful (2)

The unbridled use of the go to statement has as an immediate consequence that it becomes terribly hard to find a meaningful set of coordinates in which to describe the process progress. ... The go to statement as it stands is just too primitive, it is too much an invitation to make a mess of one's program.

goto, or structured programming constructs, are for control flow

Goto considered harmful (2)

The unbridled use of the go to statement has as an immediate consequence that it becomes terribly hard to find a meaningful set of coordinates in which to describe the process progress. ... The go to statement as it stands is just too primitive, it is too much an invitation to make a mess of one's program.

goto, or structured programming constructs, are for control flow

mutation, or function calls, are for data flow

Goto considered harmful (2)

The unbridled use of the go to statement has as an immediate consequence that it becomes terribly hard to find a meaningful set of coordinates in which to describe the process progress. ... The go to statement as it stands is just too primitive, it is too much an invitation to make a mess of one's program.

goto, or structured programming constructs, are for control flow

mutation, or function calls, are for data flow
'structured' data flow: OO

Goto considered harmful (2)

The unbridled use of the go to statement has as an immediate consequence that it becomes terribly hard to find a meaningful set of coordinates in which to describe the process progress. ... The go to statement as it stands is just too primitive, it is too much an invitation to make a mess of one's program.

goto, or structured programming constructs, are for control flow

mutation, or function calls, are for data flow
'structured' data flow: OO
yet object modifications still have effect on a distance

Goto considered harmful (2)

The unbridled use of the go to statement has as an immediate consequence that it becomes terribly hard to find a meaningful set of coordinates in which to describe the process progress. ... The go to statement as it stands is just too primitive, it is too much an invitation to make a mess of one's program.

goto, or structured programming constructs, are for control flow

mutation, or function calls, are for data flow
'structured' data flow: OO
yet object modifications still have effect on a distance

"Globals are evil"

$_, $@, $AUTOLOAD, ... -- problematic even without threads

A Gradual Change

Perl still has goto and labels; you just don't use those much
Some calculations are trivially represented by a pure function (mathematical expressions)
What this project is about, is to increase the range of situations where pure functions can be used (or existing potentially impure functions become pure with certainty)
Hence, to increase the amount of a program that's written as pure functions
you don't need to rewrite the whole of your program to have more 'purity'; make one more function pure, there's one more pure function. (Purity islands.)
'pure function' does not preclude 'method dispatch'. (Methods can be pure functions.)

Programming with functions

is not difficult to explain

again gradual: you're used to:
```
map { $_ * $_ } @x
```
there will more new constructs like 'map' and 'grep', learning about them is useful but not required (reinvent the wheel)
New tools for more benefit, e.g. type systems, are not a necessary ingredient to programming with pure functions.

It may look unusual, though.

Classification of side effects

Impure functions can read or write side effects in several ways

exert changes to the 'environment'
be subjected to changes in the 'environment'
readers and writers can be linked indirectly
the 'environment' can be outer lexical variables, global variables, mutable data structures, the world outside the process (I/O)

Classification of side effects (2)

Notable: mutable data structures can be (unintended) side effect channels.

code that doesn't exert changes can be pure
except if it's subject to changes by others

Side-effect free code and datastructures are complementing each other.

Thanks for listening!

Questions?

Get code and docs from functional-perl.org