setInterval and setTimeout patterns

gistfile1.js

// =========================
// SetInterval
// =========================

// While not truly accurate, setInterval is still fairly good, time-wise.
// Better for things like a "one second tick" but not great for expensive
// code executed at small intervals as iterations can "stack".

// (ECMAScript 5 strict mode compatible)

// First "arbitrary code" execution happens after ~1000ms, subsequent
// executions follow suit.
setInterval(function(){
  console.log(new Date()); // run some arbitrary code
}, 1000);

// First "arbitrary code" execution is synchronous, subsequent executions
// happen every ~1000ms.
setInterval((function loopy(){
  console.log(new Date()); // run some arbitrary code
  return loopy;
})(), 1000);

// =========================
// setTimeout
// =========================

// While less accurate time-wise than setInterval, setTimeout allows for
// more "breathing room" when executing expensive code at smaller intervals,
// as a next iteration will only happen ~1000ms _after_ the arbtrary code is
// executed and, not _before_ (like setInterval)

// First "arbitrary code" execution is synchronous, subsequent executions
// happen every ~1000ms.
(function loopy(){
  console.log(new Date()); // run some arbitrary code
  setTimeout( loopy, 1000 );
})();

// First "arbitrary code" execution happens after ~1000ms, subsequent
// executions follow suit.
(function loopy(){
  setTimeout(function(){
    console.log(new Date()); // run some arbitrary code
    loopy();
  }, 1000 );
})();

One caveat to the (function foo(){ ... foo(); })() pattern is that it doesn't work in BlackBerry OS 5.x or earlier, as the code inside the named function expression can't access the the function by name (for whatever reason). So, if you're writing cross-OS mobile JavaScript, you should avoid this pattern.

See this jQuery hashchange event issue for a little more information on this issue.