1. Capital^H^HolJS
brendan@mozilla.org
Sunday, September 18, 2011
2. Capital^H^HolJS
brendan@mozilla.org
• History, goals
Sunday, September 18, 2011
3. Capital^H^HolJS
brendan@mozilla.org
• History, goals
• ES6, what’s in it?
Sunday, September 18, 2011
4. Capital^H^HolJS
brendan@mozilla.org
• History, goals
• ES6, what’s in it?
• Dart: I can’t wait!
Sunday, September 18, 2011
5. Capital^H^HolJS
brendan@mozilla.org
• History, goals
• ES6, what’s in it?
• Dart: I can’t wait!
• Harmony tune-up
Sunday, September 18, 2011
6. Capital^H^HolJS
brendan@mozilla.org
• History, goals
• ES6, what’s in it?
• Dart: I can’t wait!
• Harmony tune-up
• RiverTrail demo
Sunday, September 18, 2011
7. A very brief history
mozilla
2
Sunday, September 18, 2011
8. A very brief history
• Ten days in May 1995: “Mocha”, form validation, img rollovers, scripting of Java
mozilla
2
Sunday, September 18, 2011
9. A very brief history
• Ten days in May 1995: “Mocha”, form validation, img rollovers, scripting of Java
• September 1995: “LiveScript” (did Netscape marketing go to Microsoft after?)
mozilla
2
Sunday, September 18, 2011
10. A very brief history
• Ten days in May 1995: “Mocha”, form validation, img rollovers, scripting of Java
• September 1995: “LiveScript” (did Netscape marketing go to Microsoft after?)
• December 1995: “JavaScript”, thanks to Bill Joy (a Sun Founder)
mozilla
2
Sunday, September 18, 2011
11. A very brief history
• Ten days in May 1995: “Mocha”, form validation, img rollovers, scripting of Java
• September 1995: “LiveScript” (did Netscape marketing go to Microsoft after?)
• December 1995: “JavaScript”, thanks to Bill Joy (a Sun Founder)
• 1996-1997: ES1 thanks to ECMA (now Ecma) TC39 -- see my TXJS talk
mozilla
2
Sunday, September 18, 2011
12. A very brief history
• Ten days in May 1995: “Mocha”, form validation, img rollovers, scripting of Java
• September 1995: “LiveScript” (did Netscape marketing go to Microsoft after?)
• December 1995: “JavaScript”, thanks to Bill Joy (a Sun Founder)
• 1996-1997: ES1 thanks to ECMA (now Ecma) TC39 -- see my TXJS talk
• 1999: ES3: function expressions, RegExps, try/catch/finally, switch, do-while
mozilla
2
Sunday, September 18, 2011
13. A very brief history
• Ten days in May 1995: “Mocha”, form validation, img rollovers, scripting of Java
• September 1995: “LiveScript” (did Netscape marketing go to Microsoft after?)
• December 1995: “JavaScript”, thanks to Bill Joy (a Sun Founder)
• 1996-1997: ES1 thanks to ECMA (now Ecma) TC39 -- see my TXJS talk
• 1999: ES3: function expressions, RegExps, try/catch/finally, switch, do-while
• 2005: the Ajax revolution, followed by “The Ajax Experience”
mozilla
2
Sunday, September 18, 2011
14. A very brief history
• Ten days in May 1995: “Mocha”, form validation, img rollovers, scripting of Java
• September 1995: “LiveScript” (did Netscape marketing go to Microsoft after?)
• December 1995: “JavaScript”, thanks to Bill Joy (a Sun Founder)
• 1996-1997: ES1 thanks to ECMA (now Ecma) TC39 -- see my TXJS talk
• 1999: ES3: function expressions, RegExps, try/catch/finally, switch, do-while
• 2005: the Ajax revolution, followed by “The Ajax Experience”
• 2008: ES4 RIP, Harmony founded in July at the Oslo TC39 meeting
mozilla
2
Sunday, September 18, 2011
15. A very brief history
• Ten days in May 1995: “Mocha”, form validation, img rollovers, scripting of Java
• September 1995: “LiveScript” (did Netscape marketing go to Microsoft after?)
• December 1995: “JavaScript”, thanks to Bill Joy (a Sun Founder)
• 1996-1997: ES1 thanks to ECMA (now Ecma) TC39 -- see my TXJS talk
• 1999: ES3: function expressions, RegExps, try/catch/finally, switch, do-while
• 2005: the Ajax revolution, followed by “The Ajax Experience”
• 2008: ES4 RIP, Harmony founded in July at the Oslo TC39 meeting
mozilla
• 2009: ES5: “use strict”, JSON, Object.create, etc.
2
Sunday, September 18, 2011
17. The Harmony goals
• Be a better language for writing:
mozilla
3
Sunday, September 18, 2011
18. The Harmony goals
• Be a better language for writing:
• complex applications
mozilla
3
Sunday, September 18, 2011
19. The Harmony goals
• Be a better language for writing:
• complex applications
• libraries (including the DOM) shared by those applications
mozilla
3
Sunday, September 18, 2011
20. The Harmony goals
• Be a better language for writing:
• complex applications
• libraries (including the DOM) shared by those applications
• code generators targeting the new edition
mozilla
3
Sunday, September 18, 2011
21. The Harmony goals
• Be a better language for writing:
• complex applications
• libraries (including the DOM) shared by those applications
• code generators targeting the new edition
• Better tests, if not a testable (executable) specification
mozilla
3
Sunday, September 18, 2011
22. The Harmony goals
• Be a better language for writing:
• complex applications
• libraries (including the DOM) shared by those applications
• code generators targeting the new edition
• Better tests, if not a testable (executable) specification
• Adopt de facto standards where possible
mozilla
3
Sunday, September 18, 2011
23. The Harmony goals
• Be a better language for writing:
• complex applications
• libraries (including the DOM) shared by those applications
• code generators targeting the new edition
• Better tests, if not a testable (executable) specification
• Adopt de facto standards where possible
• Keep versioning as simple and linear as possible
mozilla
3
Sunday, September 18, 2011
24. The Harmony goals
• Be a better language for writing:
• complex applications
• libraries (including the DOM) shared by those applications
• code generators targeting the new edition
• Better tests, if not a testable (executable) specification
• Adopt de facto standards where possible
• Keep versioning as simple and linear as possible
• Support a statically verifiable, object-capability secure subset mozilla
3
Sunday, September 18, 2011
26. Approved for ES6
• let, const, function in block scope
mozilla
4
Sunday, September 18, 2011
27. Approved for ES6
• let, const, function in block scope
• destructuring: let {x, y} = pt; let [s, v, o] = triple()
mozilla
4
Sunday, September 18, 2011
28. Approved for ES6
• let, const, function in block scope
• destructuring: let {x, y} = pt; let [s, v, o] = triple()
• parameter default values: function f(x, y=1, z=0) {...}
mozilla
4
Sunday, September 18, 2011
29. Approved for ES6
• let, const, function in block scope
• destructuring: let {x, y} = pt; let [s, v, o] = triple()
• parameter default values: function f(x, y=1, z=0) {...}
• rest, spread: function g(i, j, ...r) { return r.slice(i, j); }
let a = [0,1,2,3],
o = new any_constructor(...a)
mozilla
4
Sunday, September 18, 2011
30. Approved for ES6
• let, const, function in block scope
• destructuring: let {x, y} = pt; let [s, v, o] = triple()
• parameter default values: function f(x, y=1, z=0) {...}
• rest, spread: function g(i, j, ...r) { return r.slice(i, j); }
let a = [0,1,2,3],
o = new any_constructor(...a)
• proxies, weak maps: Proxy.create(handler, proto), new WeakMap
mozilla
4
Sunday, September 18, 2011
31. Approved for ES6
• let, const, function in block scope
• destructuring: let {x, y} = pt; let [s, v, o] = triple()
• parameter default values: function f(x, y=1, z=0) {...}
• rest, spread: function g(i, j, ...r) { return r.slice(i, j); }
let a = [0,1,2,3],
o = new any_constructor(...a)
• proxies, weak maps: Proxy.create(handler, proto), new WeakMap
• modules: module M { export function fast_sin(x) {...} }
mozilla
4
Sunday, September 18, 2011
32. Approved for ES6
• let, const, function in block scope
• destructuring: let {x, y} = pt; let [s, v, o] = triple()
• parameter default values: function f(x, y=1, z=0) {...}
• rest, spread: function g(i, j, ...r) { return r.slice(i, j); }
let a = [0,1,2,3],
o = new any_constructor(...a)
• proxies, weak maps: Proxy.create(handler, proto), new WeakMap
• modules: module M { export function fast_sin(x) {...} }
• iterators, generators: function* gen() { yield 1; yield 2; }
mozilla
4
Sunday, September 18, 2011
33. Approved for ES6
• let, const, function in block scope
• destructuring: let {x, y} = pt; let [s, v, o] = triple()
• parameter default values: function f(x, y=1, z=0) {...}
• rest, spread: function g(i, j, ...r) { return r.slice(i, j); }
let a = [0,1,2,3],
o = new any_constructor(...a)
• proxies, weak maps: Proxy.create(handler, proto), new WeakMap
• modules: module M { export function fast_sin(x) {...} }
• iterators, generators: function* gen() { yield 1; yield 2; }
mozilla
• comprehensions: return [a+b for a of A for b of B]
4
Sunday, September 18, 2011
40. Quasi-Literals
• Injection-safer string interpolation and domain-specific languages
mozilla
6
Sunday, September 18, 2011
41. Quasi-Literals
• Injection-safer string interpolation and domain-specific languages
• Backtick-quoted string desugars to a function call that operates on the literal
portions and substitution results:
mozilla
6
Sunday, September 18, 2011
42. Quasi-Literals
• Injection-safer string interpolation and domain-specific languages
• Backtick-quoted string desugars to a function call that operates on the literal
portions and substitution results:
• quasi`literalPart1${substitution}literalPart2` desugars to
mozilla
6
Sunday, September 18, 2011
43. Quasi-Literals
• Injection-safer string interpolation and domain-specific languages
• Backtick-quoted string desugars to a function call that operates on the literal
portions and substitution results:
• quasi`literalPart1${substitution}literalPart2` desugars to
• quasi({raw: ["literalPart1", "literalPart1"],
unescaped: ["literalPart1", "literalPart1"]},
substitution)
mozilla
6
Sunday, September 18, 2011
44. Quasi-Literals
• Injection-safer string interpolation and domain-specific languages
• Backtick-quoted string desugars to a function call that operates on the literal
portions and substitution results:
• quasi`literalPart1${substitution}literalPart2` desugars to
• quasi({raw: ["literalPart1", "literalPart1"],
unescaped: ["literalPart1", "literalPart1"]},
substitution)
• Multiline string literals w/o prefix: `literalPart1${substitution}
(yes, that was a newline!) literalPart2`
mozilla
6
Sunday, September 18, 2011
45. Quasi-Literals
• Injection-safer string interpolation and domain-specific languages
• Backtick-quoted string desugars to a function call that operates on the literal
portions and substitution results:
• quasi`literalPart1${substitution}literalPart2` desugars to
• quasi({raw: ["literalPart1", "literalPart1"],
unescaped: ["literalPart1", "literalPart1"]},
substitution)
• Multiline string literals w/o prefix: `literalPart1${substitution}
(yes, that was a newline!) literalPart2`
mozilla
• Multiline regexp literals: re`literalPart1${substitution}
(yes, that was a newline!) w+ literalPart2`
6
Sunday, September 18, 2011
46. Classes made it, barely (Yay, I think? Why am I sad?)
mozilla
7
Sunday, September 18, 2011
47. Classes made it, barely (Yay, I think? Why am I sad?)
• Sugar for prototypal object pattern, also supports closure patterns:
mozilla
7
Sunday, September 18, 2011
48. Classes made it, barely (Yay, I think? Why am I sad?)
• Sugar for prototypal object pattern, also supports closure patterns:
• class Point {
constructor(x, y) {
private x = x,
y = y;
public closed_r() { return Math.sqrt(x*x + y*y); }
}
get x() { return @x; }
get y() { return @y; }
equals(p) { return @x === p@x && @y === p@y; }
proto_r() { return Math.sqrt(@x * @x + @y * @y); }
}
mozilla
7
Sunday, September 18, 2011
49. Classes made it, barely (Yay, I think? Why am I sad?)
• Sugar for prototypal object pattern, also supports closure patterns:
• class Point {
constructor(x, y) {
private x = x,
y = y;
public closed_r() { return Math.sqrt(x*x + y*y); }
}
get x() { return @x; }
get y() { return @y; }
equals(p) { return @x === p@x && @y === p@y; }
proto_r() { return Math.sqrt(@x * @x + @y * @y); }
}
• Private member reference syntax (@x, p@x) still up in the air...
mozilla
7
Sunday, September 18, 2011
50. Classes made it, barely (Yay, I think? Why am I sad?)
• Sugar for prototypal object pattern, also supports closure patterns:
• class Point {
constructor(x, y) {
private x = x,
y = y;
public closed_r() { return Math.sqrt(x*x + y*y); }
}
get x() { return @x; }
get y() { return @y; }
equals(p) { return @x === p@x && @y === p@y; }
proto_r() { return Math.sqrt(@x * @x + @y * @y); }
}
• Private member reference syntax (@x, p@x) still up in the air...
• extends, prototype, and super work the way you want
mozilla
7
Sunday, September 18, 2011
51. Not yet in Harmony: arrow function syntax
mozilla
8
Sunday, September 18, 2011
52. Not yet in Harmony: arrow function syntax
• Arrow function syntax, instead of λ, ƒ, or # (want to save # for later)
mozilla
8
Sunday, September 18, 2011
53. Not yet in Harmony: arrow function syntax
• Arrow function syntax, instead of λ, ƒ, or # (want to save # for later)
• Just like CoffeeScript: let identity = (x) -> x
mozilla
8
Sunday, September 18, 2011
54. Not yet in Harmony: arrow function syntax
• Arrow function syntax, instead of λ, ƒ, or # (want to save # for later)
• Just like CoffeeScript: let identity = (x) -> x
• Expression body: const square = (x) -> (x * x)
mozilla
8
Sunday, September 18, 2011
55. Not yet in Harmony: arrow function syntax
• Arrow function syntax, instead of λ, ƒ, or # (want to save # for later)
• Just like CoffeeScript: let identity = (x) -> x
• Expression body: const square = (x) -> (x * x)
• Statement body: let countUsed = (str) -> {
if (str in usedWords)
usedWords[str]++;
else
usedWords[str] = 1;
}
mozilla
8
Sunday, September 18, 2011
56. Not yet in Harmony: arrow function syntax
• Arrow function syntax, instead of λ, ƒ, or # (want to save # for later)
• Just like CoffeeScript: let identity = (x) -> x
• Expression body: const square = (x) -> (x * x)
• Statement body: let countUsed = (str) -> {
if (str in usedWords)
usedWords[str]++;
else
usedWords[str] = 1;
}
• Fat arrow too: callback = (msg) => ( this.vmail.push(msg) )
mozilla
8
Sunday, September 18, 2011
57. Not yet in Harmony: arrow function syntax
• Arrow function syntax, instead of λ, ƒ, or # (want to save # for later)
• Just like CoffeeScript: let identity = (x) -> x
• Expression body: const square = (x) -> (x * x)
• Statement body: let countUsed = (str) -> {
if (str in usedWords)
usedWords[str]++;
else
usedWords[str] = 1;
}
• Fat arrow too: callback = (msg) => ( this.vmail.push(msg) )
• Binding forms: let f() -> “writable”
const K() -> “readonly” mozilla
8
Sunday, September 18, 2011
58. Not yet in Harmony: block lambda revival
• Inspired by Smalltalk via Ruby
let empty = {||}; // no need for space between bars
assert(empty() === undefined);
assert(typeof empty === "function"); // native and implements [[Call]]
assert(empty.length === 0);
let identity = {|x| x}; // completion is return value
assert(identity(42) === 42);
assert(identity.length === 1);
let a = [1, 2, 3, 4];
let b = a.map {|e| e * e} // paren-free call with block is
// idiomatic control structure so
// no semicolon at end
mozilla
print(b); // [1, 4, 9, 16]
9
Sunday, September 18, 2011
59. More block lambda revival
• Paren-free calls, control effects: you know you want it...
b = a.map {|e| // newline in block ok
e * e * e} // newline after ends call
function find_first_odd(a) {
a.forEach { |e, i|
if (e & 1) return i; } // returns from function
return -1;
}
function escape_return() {
return {|e| return e};
}
b = escape_return();
try { b(42); } catch (e) {} // error, return from
// inactive function
mozilla
10
Sunday, September 18, 2011
60. Syntax, yay
mozilla
11
Sunday, September 18, 2011
61. Syntax, yay
• Do we need shorter function syntax at all?
mozilla
11
Sunday, September 18, 2011
62. Syntax, yay
• Do we need shorter function syntax at all?
• Some say no. @mikeal: “I got 99 problems and JavaScript syntax ain’t one.”
mozilla
11
Sunday, September 18, 2011
63. Syntax, yay
• Do we need shorter function syntax at all?
• Some say no. @mikeal: “I got 99 problems and JavaScript syntax ain’t one.”
• Others say yes, but divide into roughly three camps:
mozilla
11
Sunday, September 18, 2011
64. Syntax, yay
• Do we need shorter function syntax at all?
• Some say no. @mikeal: “I got 99 problems and JavaScript syntax ain’t one.”
• Others say yes, but divide into roughly three camps:
• Arrow function syntax, as in CoffeeScript.
mozilla
11
Sunday, September 18, 2011
65. Syntax, yay
• Do we need shorter function syntax at all?
• Some say no. @mikeal: “I got 99 problems and JavaScript syntax ain’t one.”
• Others say yes, but divide into roughly three camps:
• Arrow function syntax, as in CoffeeScript.
• Block lambdas. @jashkenas: “For the record, I too favor [block lambdas] if
arrows in JS will need curlies to delimit blocks. Curlies or arrows, not both.”
mozilla
11
Sunday, September 18, 2011
66. Syntax, yay
• Do we need shorter function syntax at all?
• Some say no. @mikeal: “I got 99 problems and JavaScript syntax ain’t one.”
• Others say yes, but divide into roughly three camps:
• Arrow function syntax, as in CoffeeScript.
• Block lambdas. @jashkenas: “For the record, I too favor [block lambdas] if
arrows in JS will need curlies to delimit blocks. Curlies or arrows, not both.”
• Some leading ugly character, maybe -- unsure about this, return.
mozilla
11
Sunday, September 18, 2011
68. Dart to the heart
• Some things I’ve heard & read
Sunday, September 18, 2011
69. Dart to the heart
• Some things I’ve heard & read
• “If JavaScript does not have
{classes, “guards”}, then it will be
REPLACED!” [May 2010 TC39 meeting]
Sunday, September 18, 2011
70. Dart to the heart
• Some things I’ve heard & read
• “If JavaScript does not have
{classes, “guards”}, then it will be
REPLACED!” [May 2010 TC39 meeting]
• “better performance profile”
Sunday, September 18, 2011
71. Dart to the heart
• Some things I’ve heard & read
• “If JavaScript does not have
{classes, “guards”}, then it will be
REPLACED!” [May 2010 TC39 meeting]
• “better performance profile”
• “amenable to tooling”, “(e.g. with
optional types)”
Sunday, September 18, 2011
72. Dart to the heart
• Some things I’ve heard & read
• “If JavaScript does not have
{classes, “guards”}, then it will be
REPLACED!” [May 2010 TC39 meeting]
• “better performance profile”
• “amenable to tooling”, “(e.g. with
optional types)”
• “Fundamental problems [like the]
single Number primitive....”
Sunday, September 18, 2011
73. iOS, lolwut?
mozilla
13
Sunday, September 18, 2011
74. iOS, lolwut?
• Leaked memo: “The emergence of compelling alternative platforms like iOS has
meant that the web platform must compete on its merits, not just its reach.”
mozilla
13
Sunday, September 18, 2011
75. iOS, lolwut?
• Leaked memo: “The emergence of compelling alternative platforms like iOS has
meant that the web platform must compete on its merits, not just its reach.”
• m-k-s on Hacker News: “If they think per the leaked mail that devs are choosing
iOS over browsers because of JS, they have their heads in the sand.
Those choices are happening due to iOS as a platform not because devs are
somehow so much more happier learning/using Objective-C! The best thing that
Google could do is to KEEP working hard on improving the browser platform.”
mozilla
13
Sunday, September 18, 2011
77. Food for thought
• The “REPLACED” threat and “optional types” suggest either
mozilla
14
Sunday, September 18, 2011
78. Food for thought
• The “REPLACED” threat and “optional types” suggest either
• Optional static types a la Strongtalk [Bracha & Griswold]
mozilla
14
Sunday, September 18, 2011
79. Food for thought
• The “REPLACED” threat and “optional types” suggest either
• Optional static types a la Strongtalk [Bracha & Griswold]
• Optional runtime guards, as proposed for Harmony but not in ES6
mozilla
14
Sunday, September 18, 2011
80. Food for thought
• The “REPLACED” threat and “optional types” suggest either
• Optional static types a la Strongtalk [Bracha & Griswold]
• Optional runtime guards, as proposed for Harmony but not in ES6
• Syntax: let i :: int = 0; function f(a :: G)...
mozilla
14
Sunday, September 18, 2011
81. Food for thought
• The “REPLACED” threat and “optional types” suggest either
• Optional static types a la Strongtalk [Bracha & Griswold]
• Optional runtime guards, as proposed for Harmony but not in ES6
• Syntax: let i :: int = 0; function f(a :: G)...
• Semantics are runtime-pluggable, tie into trademarks proposal (nominal)
mozilla
14
Sunday, September 18, 2011
82. Food for thought
• The “REPLACED” threat and “optional types” suggest either
• Optional static types a la Strongtalk [Bracha & Griswold]
• Optional runtime guards, as proposed for Harmony but not in ES6
• Syntax: let i :: int = 0; function f(a :: G)...
• Semantics are runtime-pluggable, tie into trademarks proposal (nominal)
• Proposals came at last minute (May TC39 meeting), not ready for ES6
mozilla
14
Sunday, September 18, 2011
83. Food for thought
• The “REPLACED” threat and “optional types” suggest either
• Optional static types a la Strongtalk [Bracha & Griswold]
• Optional runtime guards, as proposed for Harmony but not in ES6
• Syntax: let i :: int = 0; function f(a :: G)...
• Semantics are runtime-pluggable, tie into trademarks proposal (nominal)
• Proposals came at last minute (May TC39 meeting), not ready for ES6
• Why not int32, int64, bignum, as well as number (double)? mozilla
14
Sunday, September 18, 2011
85. The trouble with numbers
• How to introduce a sane numeric tower for JS?
mozilla
15
Sunday, September 18, 2011
86. The trouble with numbers
• How to introduce a sane numeric tower for JS?
• Suffixes: 4294967295u + 1 === 0u,
2147483647i + 1 === -2147483648i; // WTF!
mozilla
15
Sunday, September 18, 2011
87. The trouble with numbers
• How to introduce a sane numeric tower for JS?
• Suffixes: 4294967295u + 1 === 0u,
2147483647i + 1 === -2147483648i; // WTF!
• Promotion rules as in C, but with dynamic types
mozilla
15
Sunday, September 18, 2011
88. The trouble with numbers
• How to introduce a sane numeric tower for JS?
• Suffixes: 4294967295u + 1 === 0u,
2147483647i + 1 === -2147483648i; // WTF!
• Promotion rules as in C, but with dynamic types
• This smells like trouble
mozilla
15
Sunday, September 18, 2011
89. The trouble with numbers
• How to introduce a sane numeric tower for JS?
• Suffixes: 4294967295u + 1 === 0u,
2147483647i + 1 === -2147483648i; // WTF!
• Promotion rules as in C, but with dynamic types
• This smells like trouble
• Better: use bignum arithmetic; use int arithmetic; etc.
mozilla
15
Sunday, September 18, 2011
91. Better performance predictability
• JS engines currently have performance predictability problems, it’s true
mozilla
16
Sunday, September 18, 2011
92. Better performance predictability
• JS engines currently have performance predictability problems, it’s true
• Speculation can go wrong, fall back on slow paths
mozilla
16
Sunday, September 18, 2011
93. Better performance predictability
• JS engines currently have performance predictability problems, it’s true
• Speculation can go wrong, fall back on slow paths
• Array holes, prototype properties being deleted, unshadowing/fore-shadowing
all hurt a bit
mozilla
16
Sunday, September 18, 2011
94. Better performance predictability
• JS engines currently have performance predictability problems, it’s true
• Speculation can go wrong, fall back on slow paths
• Array holes, prototype properties being deleted, unshadowing/fore-shadowing
all hurt a bit
• But JS engines are still improving rapidly (e.g. the SpiderMonkey Type Inference
work by Brian Hackett)
mozilla
16
Sunday, September 18, 2011
95. Better performance predictability
• JS engines currently have performance predictability problems, it’s true
• Speculation can go wrong, fall back on slow paths
• Array holes, prototype properties being deleted, unshadowing/fore-shadowing
all hurt a bit
• But JS engines are still improving rapidly (e.g. the SpiderMonkey Type Inference
work by Brian Hackett)
• It’s way too soon to throw in the towel
mozilla
16
Sunday, September 18, 2011
96. Better performance predictability
• JS engines currently have performance predictability problems, it’s true
• Speculation can go wrong, fall back on slow paths
• Array holes, prototype properties being deleted, unshadowing/fore-shadowing
all hurt a bit
• But JS engines are still improving rapidly (e.g. the SpiderMonkey Type Inference
work by Brian Hackett)
• It’s way too soon to throw in the towel
• We could add more opt-into-high-performance knobs, but: No.
mozilla
16
Sunday, September 18, 2011
98. Harmony tune-up
• It’s good to question the process
mozilla
17
Sunday, September 18, 2011
99. Harmony tune-up
• It’s good to question the process
• Harmony is both conservative (consensus is hard to achieve) and path-
dependent (because macro research takes time, we’re adding syntactic special
forms that could be macros -- if only we had macros)
mozilla
17
Sunday, September 18, 2011
100. Harmony tune-up
• It’s good to question the process
• Harmony is both conservative (consensus is hard to achieve) and path-
dependent (because macro research takes time, we’re adding syntactic special
forms that could be macros -- if only we had macros)
• Should we take the Dart clues to heart, without over-reacting?
mozilla
17
Sunday, September 18, 2011
101. Harmony tune-up
• It’s good to question the process
• Harmony is both conservative (consensus is hard to achieve) and path-
dependent (because macro research takes time, we’re adding syntactic special
forms that could be macros -- if only we had macros)
• Should we take the Dart clues to heart, without over-reacting?
• My suggestion is to prototype in SpiderMonkey and V8 some of the strawman
proposals that did not make ES6 now, and see if any deserve promotion to ES6
mozilla
17
Sunday, September 18, 2011
103. RiverTrail: Parallel JS
• A ParallelArray library, like typed arrays but immutable.
mozilla
18
Sunday, September 18, 2011
104. RiverTrail: Parallel JS
• A ParallelArray library, like typed arrays but immutable.
• map, reduce, combine, filter, partition, scan, scatter -- all produce fresh
ParallelArray results
mozilla
18
Sunday, September 18, 2011
105. RiverTrail: Parallel JS
• A ParallelArray library, like typed arrays but immutable.
• map, reduce, combine, filter, partition, scan, scatter -- all produce fresh
ParallelArray results
• Relies on commutativity to parallelize arithmetic operations (this does inject
some f.p. non-determinism).
mozilla
18
Sunday, September 18, 2011
106. RiverTrail: Parallel JS
• A ParallelArray library, like typed arrays but immutable.
• map, reduce, combine, filter, partition, scan, scatter -- all produce fresh
ParallelArray results
• Relies on commutativity to parallelize arithmetic operations (this does inject
some f.p. non-determinism).
• A Firefox add-on that runs a Narcissus-based JS-to-OpenCL compiler over
ParallelArray code
mozilla
18
Sunday, September 18, 2011
107. RiverTrail: Parallel JS
• A ParallelArray library, like typed arrays but immutable.
• map, reduce, combine, filter, partition, scan, scatter -- all produce fresh
ParallelArray results
• Relies on commutativity to parallelize arithmetic operations (this does inject
some f.p. non-determinism).
• A Firefox add-on that runs a Narcissus-based JS-to-OpenCL compiler over
ParallelArray code
• Source: github.com/RiverTrail/RiverTrail (user id should’ve been IntelLabs)
mozilla
18
Sunday, September 18, 2011
108. RiverTrail: Parallel JS
• A ParallelArray library, like typed arrays but immutable.
• map, reduce, combine, filter, partition, scan, scatter -- all produce fresh
ParallelArray results
• Relies on commutativity to parallelize arithmetic operations (this does inject
some f.p. non-determinism).
• A Firefox add-on that runs a Narcissus-based JS-to-OpenCL compiler over
ParallelArray code
• Source: github.com/RiverTrail/RiverTrail (user id should’ve been IntelLabs)
• Demo code is funny: it contains a "DO NOT use strict”; directive! mozilla
18
Sunday, September 18, 2011
109. RiverTrail demo sample code
• the ParallelArray constructor builds on typed arrays:
NBody.private.initVel = new Array(numBodies);
NBody.private.initVelTA = new Float32Array(numBodies * 4);
var initAsteroidPos = new Array(50);
for (i = 0; i < 50; i++) {
initAsteroidPos[i] = new ParallelArray([
NBody.private.posTA[i * 3 + 0], // x
NBody.private.posTA[i * 3 + 1],
NBody.private.posTA[i * 3 + 2],
]);
}
. . .
NBody.private.asteroidPos = new ParallelArray(Float32Array, initAsteroidPos);
. . .
NBody.private.vel = new ParallelArray(Float32Array, NBody.private.initVel);
mozilla
19
Sunday, September 18, 2011
110. ParallelArray methods in action
• combine method is a workhorse (takes variable number of args)
"animateTickParallel": function animateTickParallel() {
// increment (+=) velocity by acceleration
var newVel = NBody.private.vel.combine(
1,
low_precision(NBody.bodyVelocityLoopified),
NBody.private.pos,
numBodies,
NBody.Constant.deltaTime,
NBody.Constant.epsSqr,
NBody.time,
NBody.private.asteroidPos
);
. . . mozilla
20
Sunday, September 18, 2011
112. Always bet on JS
• First they said JS couldn’t be useful
for buildling “rich Internet apps”
Sunday, September 18, 2011
113. Always bet on JS
• First they said JS couldn’t be useful
for buildling “rich Internet apps”
• Then they said it couldn’t be fast
Sunday, September 18, 2011
114. Always bet on JS
• First they said JS couldn’t be useful
for buildling “rich Internet apps”
• Then they said it couldn’t be fast
• Then they said it couldn’t be fixed
Sunday, September 18, 2011
115. Always bet on JS
• First they said JS couldn’t be useful
for buildling “rich Internet apps”
• Then they said it couldn’t be fast
• Then they said it couldn’t be fixed
• Then it couldn’t do multicore/GPU
Sunday, September 18, 2011
116. Always bet on JS
• First they said JS couldn’t be useful
for buildling “rich Internet apps”
• Then they said it couldn’t be fast
• Then they said it couldn’t be fixed
• Then it couldn’t do multicore/GPU
• Wrong every time!
Sunday, September 18, 2011
117. Always bet on JS
• First they said JS couldn’t be useful
for buildling “rich Internet apps”
• Then they said it couldn’t be fast
• Then they said it couldn’t be fixed
• Then it couldn’t do multicore/GPU
• Wrong every time!
• My advice: always bet on JS
Sunday, September 18, 2011
118. Q&A
• @BrendanEich on twitter
• brendan@mozilla.org
• es-discuss@mozilla.org
mozilla
22
Sunday, September 18, 2011