Personally, I think that this process gets you down to the core of what really makes a character himself. You have no choice but to do this if your goal is to take anachronistic characters from present times like Indiana Jones and Sherlock Holmes and Batman, and place them in a context of a sword and sorcery campaign with an expectation of a rough power balance between the players.
The core of what really makes a character is going to vary from person to person. I mean, Jack Sparrow isn't much different from, say, Dionysus or Loki or Coyote if you want to reach back to near-prehistoric times. They're basically the same character, all that's changed is the genre around them. And yet a rockstar pirate scumbag is a more resonant character today than any of our olde-tyme religions.
If all you want to do is emulate "trickster" in D&D, it's not especially difficult, in any edition (thief/rogue? Go! You might even choose bard or wizard/mage/sorcerer/illusionist, depending).
The thing is, that's not always especially satisfying. Sometimes, you WANT various genre tropes because they make the archetype a little more specific. Heck, that's half the reason for a class-based system to begin with: archetypes group related abilities and give you a sensible role on which to hang.
If you want to make Jack Sparrow be true to Jack Sparrow and not have all the extra baggage that other classes carry with them, you might want your own "Pirate" class.
I have this inkling in the back of my head that, at some point, 4e will have a base "Pirate" class. My guess is that it will come with all of the genre tropes attached, and not have significant extra baggage.
He might even be a Striker, so he'll have some real similarities to the rogue and the ranger already. Maybe even share some abilities.
But in a narrow design system, which 4e appears to be, a new class for each archetype works better than trying to shoehorn the archetype into an existing class.