To create a dual weilding sword lovin' priest of Kelanen in 3E, hands down the better cleric will take TWF feat rather than 2 levels of fighter and losing an entire level of spells + 2 caster levels.
The better holy swordsman will take the 2 fighter levels (+1 BAB and an extra feat).
This is a design flaw with 3E. You have to stay cleric or appropriate PrC to be a cleric at a decent power level to ECL. The TWF feat blows chunks for a cleric because of the stat requirement. 15+ dex on a cleric? A cleric who takes TWF for a feat wasted one of his limited feats and is sub-optimal. If the cleric takes some fighter levels he has diluted his caster levels and is sub-optimal. Where is the good design? Where is the increased choice? A cleric must be a caster first -melee second or else he is sub-optimal. So whether the class design is flawed or the feat is flawed; there is a flaw that needs to be corrected.
The simple thing would be to fix the feat, but that would be errata not a new edition. 4E seems to be addressing the larger change of the class/multiclass system. I feel this is a good choice/design because although the feat system in 3E allowed many choices, it was only an illusion of choice because most feat choices were sub-optimal for classes. Tie the big feats to classes that can use them and use the 4E feats to customize those talent trees. It does help 'noob' proof the game but that is a good thing. If you're able enough to tinker with sub-optimal feat choices in 3E, I am sure you will be able to jigger the multiclass system in 4E.
I will also add that clerics (or druids) are horrible examples to use for this discussion because they bring way to much to the table to begin with. The 3E cleric is a juggernaught that needs no feats or 'build' to be effective. It is a class without peer in the PHB and a PrC amongst NPC classes when you add in splat books.