Slideshare uses cookies to improve functionality and performance, and to provide you with relevant advertising. If you continue browsing the site, you agree to the use of cookies on this website. See our User Agreement and Privacy Policy.

Slideshare uses cookies to improve functionality and performance, and to provide you with relevant advertising. If you continue browsing the site, you agree to the use of cookies on this website. See our Privacy Policy and User Agreement for details.

Like this presentation? Why not share!

13,256 views

12,633 views

12,633 views

Published on

Slides I prepared for the 25 July 2013 Ecma TC39 meeting, on Value Objects in JS, an ES7 proposal

Published in:
Technology

No Downloads

Total views

13,256

On SlideShare

0

From Embeds

0

Number of Embeds

553

Shares

0

Downloads

37

Comments

0

Likes

18

No embeds

No notes for slide

- 1. Value Objects Brendan Eich <brendan@mozilla.org> Sunday, July 28, 13
- 2. Use Cases • symbol, arguably • int64, uint64 • Int32x4, Int32x8 (SIMD) • float32 • Float32x4, Float32x8 (SIMD) • bignum • decimal • rational • complex Sunday, July 28, 13
- 3. Overloadable Operators •| ^ & •== •< <= •<< >> >>> •+ - •* / % •~ boolean-test unary- unary+ Sunday, July 28, 13
- 4. Preserving Boolean Algebra • != and ! are not overloadable to preserve identities including • X ? A : B <=> !X ? B : A • !(X && Y) <=> !X || !Y • !(X || Y) <=> !X && !Y • X != Y <=> !(X == Y) Sunday, July 28, 13
- 5. Preserving Relational Relations • > and >= are derived from < and <= as follows: • A > B <=> B < A • A >= B <=> B <= A • We provide <= in addition to < rather than derive A <= B from !(B < A) in order to allow the <= overloading to match the same value object’s == semantics -- and for special cases, e.g., unordered values (NaNs) Sunday, July 28, 13
- 6. Strict Equality Operators • The strict equality operators, === and !==, cannot be overloaded • They work on frozen-by-deﬁnition value objects via a structural recursive strict equality test • Same-object-reference remains a fast-path optimization Sunday, July 28, 13
- 7. Why Not Double Dispatch? • Left-ﬁrst asymmetry (v value, n number): • v + n ==> v.add(n) • n + v ==> v.radd(n) • Anti-modular: exhaustive other-operand type enumeration required in operator method bodies • Consequent loss of compositionality: complex and rational cannot be composed to make ratplex without modifying source or wrapping in proxies Sunday, July 28, 13
- 8. Cacheable Multimethods • Proposed in 2009 by Christian Plesner Hansen (Google) in es-discuss • Avoids double-dispatch drawbacks from last slide: binary operators implemented by 2-ary functions for each pair of types • Supports PIC optimizations (Christian was on theV8 team) • Background reading: [Chambers 1992] Sunday, July 28, 13
- 9. Binary Operator Example • For the expression v + u • Let p = v.[[Get]](@@ADD) • If p is not an Array, throw a TypeError • Let q = u.[[Get]](@@ADD_R) • If q is not an Array, throw a TypeError • Let r = p intersect q • If r.length != 1 throw a TypeError • Let f = r[0]; if f is not a function, throw • Evaluate f(v, u) and return the result Sunday, July 28, 13
- 10. API Idea from CPH 2009 function addPointAndNumber(a, b) { return Point(a.x + b, a.y + b); } Function.defineOperator('+', addPointAndNumber, Point, Number); function addNumberAndPoint(a, b) { return Point(a + b.x, a + b.y); } Function.defineOperator('+', addNumberAndPoint, Number, Point); function addPoints(a, b) { return Point(a.x + b.x, a.y + b.y); } Function.defineOperator('+', addPoints, Point, Point); Sunday, July 28, 13
- 11. Literal Syntax • int64(0) ==> 0L // as in C# • uint64(0) ==> 0UL // as in C# • float32(0) ==> 0f // as in C# • bignum(0) ==> 0I // as in F# • decimal(0) ==> 0m // or M, C/F# • We want a syntax extension mechanism, but declarative not runtime API • This suggests declarative syntax for operator deﬁnition -- and scoped usage too Sunday, July 28, 13
- 12. To new or not to new? • new connotes reference type semantics, heap allocation, mutability by default (that’s JS!) • Proposal: new int64(42) throws (for any scalar or “non-aggregate” value object) • Option: new Float32x4(a,b,c,d) makes a mutable 4-vector, but calling Float32x4(...) without new means observably immutable, so even stack allocatable (important to enable) • Alternative: always immutable, but then why allow new instead of call to “create a value” Sunday, July 28, 13
- 13. typeof travails and travesties • Invariant -- these two imply each other in JS: •typeof x == typeof y && x == y •x === y •0m == 0 && 0L == 0 means 0m == 0L (transitivity), but 0m !== 0L (different precision and radix) so typeof 0m != typeof 0L per the invariant • Usability favors typeof 0L == “int64” and typeof 0m == “decimal” anyway • Making typeof extensible requires a per-realm registry with throw-on-conﬂict Sunday, July 28, 13
- 14. 25 July 2013 TC39 Resolutions • Waldemar points out that NaN requires separately overloadable <= and < [Slide 5] • Intersection means function identity matters, so multimethods can break cross-realm [Slide 9] • Mark objects that I as bignum sufﬁx conﬂicts with complex [Slide 11] • Always throw on new -- value objects are never mutable and should not appear to be so, even if aggregate [Slide 12] • Need to work through any side channel hazard of the typeof registry [Slide 13] Sunday, July 28, 13

No public clipboards found for this slide

×
### Save the most important slides with Clipping

Clipping is a handy way to collect and organize the most important slides from a presentation. You can keep your great finds in clipboards organized around topics.

Be the first to comment