Codemash-Clojure.pdf

Clojure:
Functional
Concurrency for
the JVM
Howard M. Lewis Ship
TWD Consulting
hlship@gmail.com
1 © 2010 Howard Lewis Ship

Clojure: Why
Functional
Programming
Matters
TWD Consulting
hlship@gmail.com

Clojure: Towards
the Essence of
Programming
TWD Consulting
hlship@gmail.com

essence
noun
the intrinsic nature or indispensable quality of
something, esp. something abstract, that
determines its character : conﬂict is the essence of
drama.


Mainstream Programming

Applications

Frameworks

Libraries

Language

Operating System


Ceremony
vs.
Essence

Is your language ...


… holding
you back?

Java: Data Encapsulated in Objects
Stock Stock Stock
ticker: AAPL ticker: MSFT ticker: ORCL
lastTrade: 203.25 lastTrade: 29.12 lastTrade: 21.90
open: 204.50 open: 29.08 open: 21.83
shares: 100 shares: 50 shares: 200

public static void sortByLastTrade(List<Stock> portfolio)
{
Comparator<Stock> c = new Comparator<Stock>()
{
public int compare(Stock o1, Stock o2)
{
public static void sortByOpen(List<Stock> portfolio)
return o1.getLastTrade() - o2.getLastTrade();
{
}
Comparator<Stock> c = new Comparator<Stock>()
};
{
public int compare(Stock o1, Stock o2)
Collections.sort(portfolio, c);
{
}
return o1.getOpen() - o2.getOpen();
}
};

Collections.sort(portfolio, c);
}


Clojure: Data in Maps and Lists
:ticker AAPL :ticker MSFT :ticker ORCL

:last-trade 203.25 :last-trade 29.12 :last-trade 21.90

{ :open 204.50 } { :open 29.08 }{ :open 21.83 }
:shares 100 :shares 50 :shares 200

user=> portfolio
[{:ticker "AAPL", :last-trade 203.25M, :open 204.50M, :shares 100}
{:ticker "MSFT", :last-trade 29.12M, :open 29.08M, :shares 50}
{:ticker "ORCL", :last-trade 21.90M, :open 21.83M, :shares 200}]
user=> (sort-by :last-trade portfolio)
({:ticker "ORCL", :last-trade 21.90M, :open 21.83M, :shares 200}
{:ticker "AAPL", :last-trade 203.25M, :open 204.50M, :shares 100})
user=> (sort-by :shares portfolio)
({:ticker "MSFT", :last-trade 29.12M, :open 29.08M, :shares 50}
{:ticker "AAPL", :last-trade 203.25M, :open 204.50M, :shares 100}
{:ticker "ORCL", :last-trade 21.90M, :open 21.83M, :shares 200})
user=>


Functional JavaScript

var portfolio = [ { ticker: "AAPL", lastTrade: 203.25, open: 204.50},
{ ticker: "MSFT", lastTrade: 29.12, open: 29.08 },
{ ticker: "ORCL", lastTrade: 21.90, open: 21.83 } ]

portfolio.sortBy(function (stock) { return stock.lastTrade }).toJSON()
[{"ticker": "ORCL", "lastTrade": 21.9, "open": 21.83},
{"ticker": "MSFT", "lastTrade": 29.12, "open": 29.08},
{"ticker": "AAPL", "lastTrade": 203.25, "open": 204.5}]


Clojure: The Language


Reading Lisp

(defn render-json
"Renders JSON content (typically, a map or a seq) as the
response. The response content type is set to
"application/json". Returns true."
[env json-value]
(let [response (-> env :servlet-api :response)]
(.setContentType response "application/json")
(with-open [writer (.getWriter Aargh!
response)]
(binding [*out* writer]
(print-json json-value))))
true)


Code is Data

Quoted list of
'(1 2 3) numbers

(biggest 5 42) Function call

Function deﬁnition

(defn biggest
"Find the maximum of two numbers"
[x y]
(if (> x y) x y))


Java: Complex Structure
int result = a * b - start + offset;

Higher precendence, then left to right: ((a * b) - start) + offset

+

(+
- offset (-
(* a b)
* start start))
offset)
a b


Clojure:
Form Is
Structure

Read Eval Print Loop

user=> (defn biggest
"Find the maximum of two numbers"
[x y]
(if (> x y) x y))
#=(var user/biggest)
user=> (biggest 5 42)
42
user=> (doc biggest)
-------------------------
user/biggest
([x y])
Find the maximum of two numbers
nil
user=> '(1 2 3)
(1 2 3)
user=> '(biggest 5 42)
(biggest 5 42)
user=> (first '(biggest 5 42))
biggest
user=> (eval '(biggest 5 42))
42


REPL and Compilation

Source Code

Repl Input Clojure
User Classes Java
Evaluator
Compiler
Clojure
Source Files Java Libraries

JVM

Operating System


Clojure Literals

user=> 42
42
user=> "A Clojure String"
"A Clojure String"
user=> space
space
user=> A
A
user=> nil
nil
user=> :balance
:balance
user=> true
true
user=> false
false


Clojure Literals

user=> 5
5
user=> 5.001
5.001
user=> 22/7
22/7
user=> (* 2 22/7)
44/7
user=> (* 100000 100000 100000)
1000000000000000
user=> (+ 5. 0.000000000000000001)
5.0
user=> (+ 5.0M 0.000000000000000001M)
5.000000000000000001M


Java Interop: Function Calls
(.method-name receiver arguments…)

(.setNamespaceAware factory true)

(.. factory newSAXParser (parse src handler))

(.parse (.newSAXParser factory) src handler)

factory.newSAXParser().parse(src, handler);


Clojure Collections: Lists

4
lst
user=> (def lst '(1 2 3))
#=(var user/lst)
user=> lst
(1 2 3)
user=> (first lst)
1 1
user=> (rest lst)
(2 3)
user=> (conj lst 4) 2
(4 1 2 3)
user=> (cons 4 lst)
(4 1 2 3) 3


Clojure Collections: Vectors

user=> (def v [:moe :larry :curly])
#=(var user/v)
user=> v
[:moe :larry :curly]
user=> (first v)
:moe
user=> (rest v)
(:larry :curly)
user=> (conj v :shemp)
[:moe :larry :curly :shemp]
user=> (cons :shemp v)
(:shemp :moe :larry :curly)
user=> v
[:moe :larry :curly]
user=> (v 1)
:larry

vector is a
function of
its indexes


Clojure Collections: Maps

user=> (def m {:first-name "Howard" :last-name "Lewis Ship"})
#=(var user/m)
user=> m
{:last-name "Lewis Ship", :first-name "Howard"}
user=> (get m :last-name)
"Lewis Ship" map is a
user=> (m :last-name) function of
"Lewis Ship"
user=> (assoc m :company "TWD")
its keys
{:company "TWD", :last-name "Lewis Ship", :first-name "Howard"}
user=> m
{:last-name "Lewis Ship", :first-name "Howard"}
user=> (:first-name m)
"Howard"
user=> (:ssn m)
nil
Keywords are
functions, too!


Clojure Collections: Sets

user=> (def s #{"Howard" "Suzanne" "Molly" "Jim"})
#=(var user/s)
user=> s
#{"Howard" "Jim" "Molly" "Suzanne"}
user=> (contains? s "Howard")
true
user=> (contains? s "howard")
false set is a
user=> (s "Howard") function of
"Howard" its elements
user=> (s "Rhys")
nil
user=> (conj s "Howard")
#{"Howard" "Jim" "Molly" "Suzanne"}
user=> (conj s "Scott")
#{"Howard" "Jim" "Molly" "Suzanne" "Scott"}


Functional Programming


My First Program

No it
10 X = 1 doesn't!
20 PRINT X
30 X = X + 1
40 GOTO 20


No
Mutable
State

No Side
Effects

Purity

Simplicity
Predictability
Testability

Functional Programming in Java

public void saveOrUpdate(final Employee employee)
{
HibernateCallback callback = new HibernateCallback()
{
public Object doInHibernate(Session session)
throws HibernateException,SQLException
{
session.saveOrUpdate(employee);
return null;
}
};

hibernateTemplate.execute(callback);
}

Outer function controls the
SwingUtilities.invokeLater(new Runnable() context:
{
public void run() • Thread
{ • Exception handling
progressBar.setValue(progressBar.getValue() + 1);
} • Parameters
});

Functional Java Collections

public interface Predicate<T>
{
boolean accept(T value);
}

public static <T> Collection<T> filter(Predicate<T> pred, Collection<T> coll)
{
Collection<T> out = new ArrayList<T>();

for (T item : coll)
{
if (pred.accept(item))
out.add(item);
}

return out;
}

return CollectionUtils.filter(new Predicate<String>()
{
public boolean accept(String value)
{
return !value.startsWith("."); Essence
}
}, names);


Functional Clojure Collections
Function
Anonymous parameter
function

(filter #(not (.startsWith % ".")) names)

Member
access form

user=> (def names ["fred" "barney" ".hidden" "wilma"])
#=(var user/names)
user=> (filter #(not (.startsWith % ".")) names)
("fred" "barney" "wilma")
user=> (remove #(.startsWith % ".") names)
("fred" "barney" "wilma")
user=>


Composing Functions
(filter #(not (.startsWith % ".")) names)

function as
parameter to
function

(defn require-extension [ext]
!Closure
Oriented
(fn [file-name]
(= ext (last (split-string file-name ".")))))

Programming"
function as
return value

(filter (require-extension "gz") names)

composing functions

Java: Iterative Steps


Java: Data Encapsulated in
Objects
Stock Stock Stock
ticker: AAPL ticker: MSFT ticker: ORCL
lastTrade: 203.25 lastTrade: 29.12 lastTrade: 21.90
open: 204.50 open: 29.08 open: 21.83
shares: 100 shares: 50 shares: 200

public static List<Double> getOpens(List<Stock> portfolio)
{
List<Double> result = new ArrayList<Double>();

for (Stock stock : portfolio)
{
result.add(stock.getOpen());
}

return result;
}


Clojure: Flow of Transformations


Clojure: Data in Transformable
Collections
:ticker AAPL :ticker MSFT :ticker ORCL

:last-trade 203.25 :last-trade 29.12 :last-trade 21.90

{ :open 204.50 } { :open 29.08 }{ :open 21.83 }
:shares 100 :shares 50 :shares 200

user=> portfolio
[{:ticker "AAPL", :last-trade 203.25M, :open 204.50M, :shares 100}
{:ticker "ORCL", :last-trade 21.90M, :open 21.83M, :shares 200}]
user=> (map :open portfolio)
(204.50M 29.08M 21.83M)
user=> (apply + (map #(* (:last-trade %) (:shares %)) portfolio))
26161.00M
user=> (map #(assoc % :delta (- (% :last-trade) (% :open))) portfolio)
({:delta -1.25M, :ticker "AAPL", :last-trade 203.25M, :open 204.50M, :shares 100}
{:delta 0.04M, :ticker "MSFT", :last-trade 29.12M, :open 29.08M, :shares 50}
{:delta 0.07M, :ticker "ORCL", :last-trade 21.90M, :open 21.83M, :shares 200})
user=>


map

La
z y
f

:open acts as a function

user=> (map :open portfolio)
(204.5 29.08 21.83)
user=> (defn last-trade-value
[stock]
(* (:last-trade stock) (:shares stock)))
#'user/last-trade-value
user=> (map last-trade-value portfolio)
(20325.00M 1456.00M 4380.00M)
user=>


Laziness is
a Virtue

user=> (take 20 (iterate inc 1))
(1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20)
user=> (take 20 (map * (iterate inc 1) (iterate inc 1)))
(1 4 9 16 25 36 49 64 81 100 121 144 169 196 225 256 289 324 361 400)


map

La
z y
f

N seqs # N parameters

user=> (map #(assoc %1 :sort-index %2)
(reverse (sort-by :last-trade portfolio))
(iterate inc 0))
({:sort-index 0, :ticker "AAPL", :last-trade 203.25M, :open 204.50M, :shares 100}
{:sort-index 1, :ticker "MSFT", :last-trade 29.12M, :open 29.08M, :shares 50}
{:sort-index 2, :ticker "ORCL", :last-trade 21.90M, :open 21.83M, :shares 200})
user=>


reduce

f
user=> (map #(* (% :last-trade) (% :shares)) portfolio)
(20325.00M 1456.00M 4380.00M)
user=> (reduce + (map #(* (% :last-trade) (% :shares)) portfolio))
26161.00M
user=>

(reduce + (map #(* (% :last-trade) (% :shares)) portfolio))

(reduce + '(20325.00M 1456.00M 4380.00M))

(+ (+ 20325.00M 1456.00M) 4380.00M)

(+ 21781.00M 4380.00M)

26161.00M


reduce

f
user=> (def input-string "Clojure is a fascinating language with unique
capabilities and total integration with Java.")
#'user/input-string
user=> (seq input-string)
(C l o j u r e space i s space a space f a s c i n a
t i n g space l a n g u a g e space w i t h space u
n i q u e space c a p a b i l i t i e s space a n d
space t o t a l space i n t e g r a t i o n space w
i t h space J a v a .)
user=> (reduce
(fn [m k] (update-in m [k] #(inc (or % 0))))
Optional initial value {}
(seq input-string))
{space 12, a 12, b 1, C 1, c 2, d 1, e 5, f 1, g 4, h 2, i 11,
J 1, j 1, l 4, . 1, n 7, o 3, p 1, q 1, r 2, s 3, t 8, u 4,
v 1, w 2}
user=>


ﬁlter / remove

La
z y
f
?
user=> (remove #(< 100 (% :shares)) portfolio)
({:ticker "AAPL", :last-trade 203.25, :open 204.5, :shares 100}
{:ticker "ORCL", :last-trade 21.9, :open 21.83, :shares 57})
user=> (filter #(< 100 (% :shares)) portfolio)
({:ticker "MSFT", :last-trade 29.12, :open 29.08, :shares 125})
user=>


for: list comprehension

La
z y
user=> (for [suit [:hearts :clubs :spades :diamonds]
value (range 1 4)]
[suit value])
([:hearts 1] [:hearts 2] [:hearts 3]
[:clubs 1] [:clubs 2] [:clubs 3]
[:spades 1] [:spades 2] [:spades 3]
[:diamonds 1] [:diamonds 2] [:diamonds 3])
user=> (for [x (range 0 4)
y (range 0 (inc x))]
[x y])
([0 0]
[1 0] [1 1]
[2 0] [2 1] [2 2]
[3 0] [3 1] [3 2] [3 3])
user=> (for [x (range 0 9) :when (odd? x)
y (range 1 (inc x))]
[x y])
([1 1]
[3 1] [3 2] [3 3]
[5 1] [5 2] [5 3] [5 4] [5 5]
[7 1] [7 2] [7 3] [7 4] [7 5] [7 6] [7 7])
user=>


!Somehow the idea of reusability
got attached to object-oriented
programming in the 1980s, and
no amount of evidence to the
contrary seems to be able to
shake it free."

Paul Graham
© 2010 Howard Lewis Ship

Language Ownership


Who Owns the Java Language?

James Gosling Mark Reinhold


Not You

Control of the Compiler

Source Code

Repl Input Clojure
User Classes Java
Evaluator
Compiler
Clojure
Source Files Java Libraries

JVM

Operating System


Who Owns Clojure?

Rich Hickey


Clojure: written in Clojure


Clojure Macros
short circuit at
ﬁrst false/nil
Reader

(and a b c d)

Evaluator macro expansion
and macro

Bytecode (let [and__4422__auto__ a]
(if and__4422__auto__
Generation (and b c d)
and__4422__auto__))

Recursively expanded


Macros # Special Forms
Code Forms

Macro
Expansion

def if let fn . …

Bytecode
Generation


Macros are Special Functions
(defmacro and
"Evaluates exprs one at a time, from left to right. If a form
returns logical false (nil or false), and returns that value and
doesn't evaluate any of the other expressions, otherwise it returns
the value of the last expr. (and) returns true."
([] true)
([x] x)
([x & next]
`(let [and# ~x]
(if and# (and ~@next) and#))))

• `(…) — Syntax Quote (allowing replacements)

• and# — Generate a new unique symbol

• ~x — Unquote x

• ~@next — Unquote next and splice in multiple values


Macros are Special Functions
(defmacro and
([] true)
([x] x)
(and a b c d)
([x & next]
`(let [and# ~x]
(if and# (and ~@next) and#))))

•x#a
• next # '(b c d)
• Inside syntax quote:
• and# # and__4422__auto__
• ~x # a
• ~next # '(b c d)
• (and ~@next) # (and b c d)


Simplifying Boilerplate Code
public void testLink()
{
IMocksControl control = EasyMock.createControl();
HttpServletRequest request = control.newMock(HttpServletRequest.class);
HttpServletResponse response = control.newMock(HttpServletResponse.class);

EasyMock.expect(request.getContextPath()).andReturn("/ctx");
EasyMock.expect(response.encodeURL("/ctx/accounts/list")).andReturn("*encoded*");

control.replay();

assertEquals(…, "*encoded*");

control.verify();
}

(deftest test-link
(with-mocks [request HttpServletRequest
response HttpServletResponse]
(:train
(expect .getContextPath request "/ctx")
(expect .encodeURL response "/ctx/accounts/list" "*encoded*"))
(:test
(is (= (link request response list-accounts-with-loop)
"*encoded*")))))


Embedded DSLs

(defview root-index
[env]
:html [
:head [
:title [ "Cascade Blog" ]
]
:body [
:h1 [ "Cascade Blog" ]

:html
:ul { :class "recent-postings" } [
(template-for [posting (recent-postings env)]
:li [
(render-link env show-posting (posting :id) (posting :title))
]) :head :body
]
]
])
:title :h1 :ul

"Cascade Blog" "Cascade Blog" (template-for …)


Macro Expansions
(defn list-items [coll]
(template
(format "%d items" (count coll))
:ul {:class :item-list} [
(template-for [item coll] :li [item])))

Expand simple Extend Clojure
placeholder to language from
executable code within Clojure

(defn list-items [coll]
(cascade.internal.viewbuilder/combine
(format "%d items" (count coll))
(cascade.dom/element-node :ul {:class :item-list}
(cascade.internal.viewbuilder/combine
(for [item coll]
(cascade.dom/element-node :li nil
(cascade.internal.viewbuilder/combine item)))))))


!More than anything else, I think it is the
ability of Lisp programs to manipulate
Lisp expressions that sets Lisp apart …
when I hear people complain about Lisp's
parentheses, it sounds to my ears like
someone saying: "I tried one of those
bananas, which you say are so delicious.
The white part was ok, but the yellow
part was very tough and tasted awful.""

Paul Graham
© 2010 Howard Lewis Ship

Wrap Up


Clojure
• 1.1 release: Dec 31 2009

• Simple, regular syntax

• Improves on Lisp: vectors, maps, sets
http://www.clojure.org
• Fully integrates with Java

• Impressive functional & concurrency support

• Many features not covered here


Stuart Halloway

Pragmatic Bookshelf

http://pragprog.com/titles/shcloj/programming-clojure


http://java.ociweb.com/mark/clojure/article.html


http://tapestryjava.blogspot.com


Object Oriented


Functional


Image Credits
© 2007 Jon Fife
http://flickr.com/photos/good-karma/577632972/
© 2007 Casey Marshall
http://www.flickr.com/photos/rsdio/497112391/
© 2009 Howard M. Lewis Ship
http://www.flickr.com/photos/hlship/3603090614/
© 2009 Andrew Baird
http://www.flickr.com/photos/scruffy/3708615414/
© 2008 Miles Sabin
http://www.flickr.com/photos/montpelier/2915114545/
© 2003 A. Lipson
http://www.andrewlipson.com/escher/relativity.html
© 2007 Alan Chia
http://flickr.com/photos/seven13avenue/2080281038/
© 2007 Woodley Wonderworks
http://flickr.com/photos/wwworks/2222523486/
© Randall Munroe
http://xkcd.com/297/
© 2008 Manu Gómez
http://www.flickr.com/photos/manugomi/2884678938/
© 2008 Marcin Wichary
http://www.flickr.com/photos/mwichary/2827326852/
© 2006 Marvin (PA)
http://www.flickr.com/photos/mscolly/145052885/


Image Credits
© 2007 John Kannenberg
http://www.flickr.com/photos/jkannenberg/541057337/
© 2006 scott ogilvie
http://www.flickr.com/photos/scottog/100582274/
© 2008 Ariel H.
http://www.flickr.com/photos/fotosrotas/2730733412/


Codemash-Clojure.pdf

Recomendados

Recomendados

Más contenido relacionado

La actualidad más candente

La actualidad más candente (20)

Destacado

Destacado (16)

Similar a Codemash-Clojure.pdf

Similar a Codemash-Clojure.pdf (20)

Más de Howard Lewis Ship

Más de Howard Lewis Ship (10)

Codemash-Clojure.pdf