So is
public var user:Object = {};
user["firstName"] = "Bill";
user["lastName"] = "Co开发者_如何学运维sby";
slower than if I have a value object like User?
var user:User = new User();
user.firstName = "Bill";
user.lastName = "Cosby";
public function speedTest():void
{
var typedObjects:Array = [];
var dynamicObjects:Array = [];
var typedObject:User;
var dynamicObject:Object;
var i:int = 0;
var n:int = 10000;
for (i; i < n; i++)
{
dynamicObject = {};
dynamicObjects.push(dynamicObject);
typedObject = new User();
typedObjects.push(typedObject);
}
// typed
var time:int = getTimer();
i = 0;
n = 10000;
for (i; i < n; i++)
{
typedObject = typedObjects[i];
typedObject.firstName = "Bill";
typedObject.lastName = "Cosby";
}
var end:int = getTimer() - time;
trace("TypedObject Time: ", end/1000, "sec");
// untyped
time = getTimer();
i = 0;
n = 10000;
for (i; i < n; i++)
{
dynamicObject = dynamicObjects[i];
dynamicObject["firstName"] = "Bill";
dynamicObject["lastName"] = "Cosby";
}
end = getTimer() - time;
trace("DynamicObject Time: ", end/1000, "sec");
}
Simple speed test shows that TypedObject (User
) is roughly 3-3.5x faster than UntypedObject ({}
), but it's nothing you'd ever notice in the day-to-day. Try out Grant Skinner's Performance Test Harness to run some better/advanced tests :).
3 simple trial runs:
TypedObject Time: 0.002 sec
DynamicObject Time: 0.007 sec
TypedObject Time: 0.002 sec
DynamicObject Time: 0.006 sec
TypedObject Time: 0.002 sec
DynamicObject Time: 0.006 sec
Verdict: Typed is faster than Untyped.
More importantly than any speed benefits, strong typing gives you compile-time type checking so you don't typo property names - saving a lot of developer time, which is far more expensive than clock cycles.
Though yes, static typing does carry speed benefits as well.
精彩评论