I actually waded through the posts since my last one, and all I got was a slight fever (actually, that's not from reading, but I wanted to add some drama... )
Is this for of multiclassing more dipping then anything else?
Possibly.
Let's be honest, multiclassing in 3E also was often similar to "dipping" - pick a level of Rogue for skills and sneak attack, or a level of Fighter for a bonus feats and heavy armor proficiency, or a level of Barbarian for Rage. In all these examples, you still got to improve your "core" abilities - BAB, Saves, HD. You couldn't do this for spellcasting, because for some reason, no one in 3.0 or 3.5 found a way to let BAB, spells per day, spells known or caster level stack, and the fix where Prestige Classes. (Eldritch Knight, Mystic Theurge...)
Many multiclassed characters also used this dipping to multiclass in PrCs.
4E has two design elements that need to be considered:
- Frontloading of classes, so you feel like a member of the class at level 1
- Every character is like a spellcaster
Maybe there would have been a different approach, but this is the approach 4E did, and at this point, you'll have to decide how to handle multiclassing with these elements.
3E style multiclasing is overpowered for front-loading, but fails in regards to spellcasting.
So, how do you balance each aspect?
- Reduce the benefits of the "front-loadedness" for multiclassing. (This was first done in Starwars Saga Edition). You get only a subset of the normal 1st level abilities.
- Let the "caster levels" stack, but don't add more spells/powers per day for multiclassing. You can choose powers from an expanded list, but still can pick the highest level powers.
The question is - what's the cost for this expansion? Should there be a cost at all? There is already the opportunity cost, after all?
3E multiclassing did rarely cost something for non-spellcasters. You usually got the most interesting stuff with 1-4 levels, and didn't really lose something. So this might be a bad basis for cost determination.
Starwars Saga Edition might come closer - if you multiclass from Noble into Soldier, you don' get that much. If you really want to "Soldier" more, you will need to expend feats and skill training feats to get Soldier-abilities, which means not only do you not advance in your Noble talents (that's expected opportunity cost), but you also pay with feats.
4E takes a similar approach. You pay the opportunity cost, and you pay with some feats. Is it worth it? That might depend on what you select.
A Fighter with an Ice Wall power doesn't sound so bad. First round, mark a foe, second round, put a wall between you and him - now he either has to select different targets (taking the -2 penalty) or waste some time walking around the wall. (Oh, and he will take some damage, too.) A Cleric with the Rogue's tumble power can safely reach a comrade to heal him. At this point, we don't know what other Cleric or Fighter powers might do, but I think this definitely shows that there are situations in where such powers seem useful enough. A lot of the powers usability might depend on party composition and a specific tactical situation.
...
There are, off course, other concerns to multiclassing. It's not always about power, it can also be about making the character concept work you had in mind, and organic growth. A Rogue constantly forced to fight in the front-lines might multiclass into Fighter in 3E, and wear a heavier armor. A 4E Rogue could now take a multiclass feat for Fighter, or pick a heavier armor profiency feat. The "trouble" might begin if the Rogue in our example now wants to pick up a third class - maybe he's travelling a lot through the wilderness, and would find tracking and archery more useful. In 3E, the solution would probably be to multiclass into Ranger. In 4E, he couldn't just multiclass into Ranger if he already multiclassed into Fighter. But on the other hand, he sure could pick up bow profiency or skill training in Nature.
An interesting "feature" here is: The character in 4E can retrain per default. If he sees that he doesn't really benefit from his fightery abilities any more, he could just choose to retrain his fighter training feats or his armor profiency, and pick up the Ranger multiclass feats or skill training in Nature.
Now, some people will undoubtly argue that they don't like retraining. Organic growth (to them) means that I can still see what the character did half a year or 3 levels ago. Well, that's a matter of taste, but I don't see much wrong with the idea of people losing their aptitude in a certain area, and regaining it in a new area. Sure, it can feel a bit awkward, if all it takes is one level (which could be gained in two days) to "forget" heavy armor profiency, but then, it also takes only the same level and two days to gain heavy armor proficiency, or learn two new spells (the latter 3E only

)
If it was just me, I'd consider saying that you can could take a "Vancian" approach to feats and powers you have, and you can choose each day (or a larger, arbitrary unit of time) to regain an ability you previously knew and chose to relearn. But that's house rule territory (as far as we know), so it doesn't exactly help the discussion.
---
On the now seemingly no longer that interesting topic on the profileration of more core/base classes: D&D is a class-based game. There is a limit to what a class can represent. There is a continuum on how strictly defined a class is. 3E tended to have a mix of class definitions - each spellcaster was tightly defined, as a result, multiclassing them was hard. Each non-spellcaster class was a lot more losely defined, the synergy between them was higher and multiclassing was easier with them. Prestige Classes basically where either trying to tighten a class focus stronger (you're not just a Fighter, but a Weapon Specialist, or you're not just a WIzard, but a Loremaster), or expanding them (Mystic Theurge the most prominent example)
4E seems to define classes more strictly. Paragon Paths seem to enforce this further, while multiclassing expands a little. But in the end, instead of expansion through PrCs or multiclassing, we will get new core classes. This has the advantage that you start in your desired class from day and level one, and don't get there at a later point. This has its appeal to me... (Even though I used to be more a fan of class-less games...)