Spoiler
A lot of really interesting designs in sci-fi get their kick from the rear silhouettes being unique, but I agree it's hard sometimes to find the right balance, and it's one of the harder things to get right; while there are lots of designs where they're practically mirrors (X-Wings, Star Destroyers, practically every race in Star Trek) there are the exceptional designs where the asymmetry there really gives it a kick. While it's hard to pull off, it looks really nice when it happens.
It think it's cool, when it works, because it agrees with physical reality or at least feels right. It's also one of the reasons why asymmetry in designs of this sort are risky- it's really hard to get that figured out sometimes.
IRL, an asymmetric design creates a lot of practical issues, in terms of dealing with center of mass (http://en.wikipedia.org/wiki/Center_of_mass) and more importantly, the amount of force being applied to the structure at the point where thrust is being created. There is torque to consider, as well; you may need X kilotons of force to cease rotation in Y time, but you're limited structurally and may not be able to solve for Y without shearing pins all over the place, heh.
With computerized fly-by wire and adjustable thrust levels, symmetry isn't necessary; a ship can calibrate itself and find its center of mass pretty easily and adjust accordingly (and they need to, if they're moving much mass around internally), but balanced forces are pretty important, especially if changing vectors.
With a simple rocket with complete symmetry and a center of mass that's completely centered, you're looking at a pretty simple picture- rear thrusters for major changes in the vector of travel and small thrusters in the front and perhaps in the back to change the vector the ship is pointed at at any given moment without imparting a lot of velocity.
But when the mass is distributed much more asymmetrically, this gets a lot more complicated, especially when we're talking about structures needing to handle the loads along the primary axis of thrust. A traditional rocket is very simple that way, since it just needs to be braced for 3 force vectors in two pretty specific places. Something that's basically symmetrical like the Millennium Falcon is a minor change in thrust levels to compensate. The cockpit is a large part of it being a brilliant design; the original concepts weren't terribly interesting until that was added. But in terms of mass, it's pretty minor, and the engines don't need to reflect that, although it would have been cooler if they had.
A B-Wing? Why does that work, though? It wasn't a happy accident; it's because the concept artist was thinking about the physics involved.
The reason the design works, imo, is that it's putting forces roughly where they'd make sense; there are only major thrusters right where one presumes the center of mass would be (http://img3.wikia.nocookie.net/__cb20080929233321/starwars/images/f/f2/Bwing_egvv.jpg). It feels right, because mass feels about even from the center of the thruster pods to everywhere else; we can imagine this object rotating in all directions without much distress.
It'd work even better if there were thrusters out on the end of the main wing, imo, to provide for precession, but Star Wars tech doesn't involve thrusters except where they feel like it, apparently. But it's quite clear from the early concepts (http://1.bp.blogspot.com/-Fo_9U2UK18w/UomUj0SGy-I/AAAAAAAARnM/YSc9aHVU1m4/s640/b-wings-design-art.jpg) that this concept of centering the mass was key to the development of the final form.
Just looking at other folks trying to make this kind of thing work (https://www.flickr.com/photos/46304451@N04/11329002674/) (btw, this comes from a whole series of Lego concepts (https://www.flickr.com/groups/asymmetric/pool/48553964@N04/) I think you might find interesting) the ones that are really radical and work are almost always ones where the center of mass is clearly defined by the design and the design reflects the physics to some degree.
The physics of making a working thing create a weird beauty, in short; a design where the asymmetry is oriented around the greebly bits will feel less "right" than one where it's obviously about the center of mass.
Anyhow, that was what I was getting at there, but I didn't want to step on your toes while you were figuring out the design :)
I'll probably do some drawings for this stuff at some point, but I'm still busy working on code for other stuff atm. Alex, one other thought; figuring the center of mass is wherever the ship's centroid is placed, you could have the engine length / size code auto-adjust, asymmetrically, according to the current rate of turn and the distance from the centerline of thrust; that'd look really, really cool on ships with engines asymmetrically placed and it'd be pretty easy to add in :)