As far as how you use them, there's no difference. However, from a performance standpoint the former method would be preferable. Let's extend your example a little bit:
var prototypePony1 = new Pony();
var prototypePony2 = new Pony();
var thisPony1 = new Pony2();
var thisPony2 = new Pony2();
prototypePony1.hasOwnProperty('bite'); //returns false
prototypePony2.hasOwnProperty('bite'); //returns false
thisPony1.hasOwnProperty('bite'); //returns true
thisPony2.hasOwnProperty('bite'); //returns true
Pony.prototype.bite = function() { alert('Nomnomnom!'); };
Pony2.prototype.bite = function() { alert('Nomnomnom!'); };
prototypePony1.bite(); //alerts 'Nomnomnom!'
prototypePony2.bite(); //alerts 'Nomnomnom!'
thisPony1.bite(); //alerts 'Chomp!', instance property is accessed first
delete thisPony2.bite;
thisPony2.hasOwnProperty('bite'); //returns false
thisPony2.bite(); //alerts 'Nomnomnom!'
In the example above, thisPony1
and thisPony2
both get their own copy of the bite
function, since it was defined using this
However, prototypePony1
and prototypePony2
both share the same copy of bite
from Pony
's constructor.
Once we define the bite
prototype on Pony2, the instance property is still accessed first on thisPony1
. It's not until we delete the instance property that we see the newly defined prototype property on thisPony2
.
For more detailed info on defining object methods, have a look here.