Clojure night

Clojure
Community
Night
Aria Haghighi
CTO
@aria42
Saturday, October 5, 13

Eng @ Prismatic

Eng @ Prismatic
All Clojure in the
back

Eng @ Prismatic
All Clojure in the
back
Some Clojure in
the front

<31. Why

<31. Why
2.
Open
Source

<31. Why
2.
Open
Source
3.
What’s
needed

Why Clojure
<3

Programming to Behavior

Encapsulate
The
Object

Interface
Encapsulate
The
Object

At it’s best,
OOP is about
coding to
behavior, not
state or identity

Programming to Data

Programming to Data
Inherent data
structures can
be used to
derive the
structure of a
program

Programming to Data
Ins & Outs

Programming to Data
Ins & Outs
Represent

Programming to Data
Ins & Outs
Represent
Transform

The killer feature
of Clojure isn’t that
it’s a functional
language, but the
best data-oriented
one
Programming to Data

Data Behavior

Data Behavior
sequences
maps
defn
destructuring
macros

Data Behavior
sequences
maps
defn
destructuring
macros
deftype
protocol
reify

Data Behavior
sequences
maps
defn
destructuring
macros
records
multi-
methods
deftype
protocol
reify

Data Behavior+
The crown of
Clojure’s design is
the clean
separation of data
and behavior

Open Source

Plumbing and Graph
data-oriented
composition and
dependency
prismatic/plumbing

Graph: Overview

Graph: Overview
Univariate statistics: Given
data as double[] xs

Graph: Overview
data as double[] xs
m =
1
n
nX
i=1
xi
m2 =
1
n
nX
i=1
x2
i
v = m2 m2

Graph: Overview
data as double[] xs
m =
1
n
nX
i=1
xi
m2 =
1
n
nX
i=1
x2
i
v = m2 m2
xs
n
m2m
v

Graph: Overview
data as double[] xs
xs
n
m2m
v

Graph: Overview
data as double[] xs
xs
n
m2m
v
(defn stats [{:keys [xs]}]
(let [n (count xs)
m (/ (sum xs) n)
m2 (/ (sum sq xs) n)
v (- m2 (* m m))]
{:n n :m m :m2 m2 :v v}))

The Monster let
(defn start [{:keys [a,z]}]
(let [s1 (store a ...)
s2 (store b ...)
db (sql-db c)
t2 (cron s2 db...)
...
srv (server ...)]
(fn shutdown []
(.stop srv)
...
(.flush s1))))

The Monster let
dozens of
parameters,
not compositional
s2 (store b ...)
db (sql-db c)
t2 (cron s2 db...)
...
srv (server ...)]
(fn shutdown []
(.stop srv)
...
(.flush s1))))

The Monster let
dozens of
parameters,
not compositional
structure of
(de)composition
is locked up in an
opaque function
s2 (store b ...)
db (sql-db c)
t2 (cron s2 db...)
...
srv (server ...)]
(fn shutdown []
(.stop srv)
...
(.flush s1))))

From let to Graph
(let [n (count xs)
m (/ (sum xs) n)
v (- m2 (* m m))]
{:n n :m m :m2 m2 :v v}))

From let to Graph
(let [n (count xs)
m (/ (sum xs) n)
v (- m2 (* m m))]
{:n n :m m :m2 m2 :v v}))
xs
n
m2m
v

From let to Graph
(let [n (count xs)
m (/ (sum xs) n)
v (- m2 (* m m))]
{:n n :m m :m2 m2 :v v}))
{:n (fn [xs] (count xs))
:m (fn [xs n] (/ (sum xs) n))
:m2 (fn [xs n] (/ (sum sq xs) n))
:v (fn [m m2] (- m2 (* m m)))}
xs
n
m2m
v

From let to Graph
(let [n (count xs)
m (/ (sum xs) n)
v (- m2 (* m m))]
{:n n :m m :m2 m2 :v v}))
{:n (fn [xs] (count xs))
:m (fn [xs n] (/ (sum xs) n))
:m2 (fn [xs n] (/ (sum sq xs) n))
:v (fn [m m2] (- m2 (* m m)))}
k
k
k
k
xs
n
m2m
v

Example:API Service

Example:API Service
(def api-service
(service
{:service-name “api”
:backend-port 42424
:server-threads 100}
{:store1 (instance store
{:type :s3 ...})
:memo (fnk [store1]
{:resource ...})
...
:api-server (...)}))

sane DOM
templating and
manipulation
prismatic/dommy
Dommy

fast primitive
array math
prismatic/hiphip
HipHip (Array)
w · x =
nX
i=1
wixi

HipHip Examples

HipHip Examples
Dot Product w · x =
nX
i=1
wixi

HipHip Examples
nX
i=1
wixi
Inner loop in
machine learning

HipHip Examples
nX
i=1
wixi
Inner loop in
machine learning
arg max
`
w · x`
prediction

HipHip Examples
nX
i=1
wixi
Inner loop in
machine learning
arg max
`
w · x`
prediction
P(x; w) / exp{w · x}
training

nX
i=1
wixi
HipHip Examples

nX
i=1
wixi
(defn dot-product [^doubles ws ^doubles xs]
(areduce ws idx sum 0.0
(+ sum (* (aget ws idx)
(aget xs idx))))
HipHip Examples

nX
i=1
wixi
double dotProd(double[] ws, double[] xs) {
double sum = 0.0;
for (int i=0; i < xs.length; ++i) {
sum += ws[i] * xs[i];
}
return sum;
}
HipHip Examples

nX
i=1
wixi
(defn dot-product [ws xs]
(hiphip.double/asum [w ws x xs] (* w x)))

HipHip Examples

declarative data
description and
validation
prismatic/schema
Schema

What Clojure needs

Where’s my debugger!?!

MOAR Compiler Speed
(require 'api.deploy)
go get a coffee, it’s
going to be a bit

Compiler Tooling
real IDE tools
Make Paredit a library

Multi-project support
in Leiningen

Native Client Story

Thanks
Questions?

Clojure night

More Related Content

What's hot

Viewers also liked

Similar to Clojure night

Recently uploaded

Clojure night