In my most recent javascript program (which is mostly for fun and proof-of-concept than anything else) I have a lot of different kinds of objects and of each kind I\'d have
In terms of performance, the difference comes when you add methods. If you use object literals each object needs to have a field for each method:
obj1--> { x: 10,
f1: method1,
f2: method2 }
obj2--> { x: 17,
f1: method1,
f2: method2 }
With classes, you can share common properties behind a shared prototype:
obj1--> { x:10,
__proto__: --------> { f1: method1,
} /----> f2: method2 }
|
obj2--> { x:17, |
__proto__: ---/
}
That said, the performance differences are only going to matter if you instantiate a lot of objects and those objects have many methods and many of those methods are closures. If I were you I would give a greater emphasis to code style issues: for example, with the object literal method you can use closures to simulate private variables while if the methods are in a shared public prototype then all your instance variables need to be public.