Question

I’m relatively new to JavaScript and repeatedly find myself writing methods in a helper object which take in a callback as a parameter e.g.

var utilities = {
    getTweets: function (user, maxTweets, callBack) {
            var obj = $(this);
            $.getJSON('http://api.twitter.com/1/statuses/user_timeline.json?callback=?&screen_name=' + user + "&count=" + maxTweets, function (data) {
                callBack(data);
            });
    };

I then call it like so:

 utilities.getTweets("TESTUSER", 4, function (tweets) {
                .....
             });

Given I am calling the code above using setInterval is this likely to leak over time/is there a better way to write this?

Answer 1

What you’re doing is mostly fine, except that there’s no need to create the extra closure. Passing a closure written like:

function (data) {
      callBack(data);
}

is just the same as passing callBack directly in the parameter list.

However if you can guarantee running with jQuery 1.5 or later, then a better method is to just have getTweets() return the JQXHR object, and then you can use “deferred” methods in the client code:

var utilities = {
    getTweets: function (user, maxTweets) {
        var uri = 'http://api.twitter.com/1/statuses/user_timeline.json?callback=?';
        var data = {
            screen_name: user,
            count: maxTweets
        };
        return $.getJSON(uri, data);
    });
};

and then in the client code:

utilities.getTweets(user, maxTweets).done( /* your callback here */ );

In this way you can completely decouple the callback from the implementation. Indeed you can register multiple callbacks, and error handlers, all without touching the implementation of utilities.

NB: use of a map for data above also protects your code against parameter injection.