Not at all.
Fair enough.
I'm not really able to compare 4e to 3E in this respect, as I don't have enough experience with 3E. Likewise for full bells-and-whistles 2nd ed AD&D.
When I compare 4e in this respect to Basic D&D and 1st ed AD&D, I would see the biggest changes of this sort are in respect to wizards, rangers and druids. Rangers and druids in AD&D are quirky enough (and arguably overpowerd enough) that the changes are tolerable - one set of oddities is replaced by another, and rangers - especially archer rangers - make for an easy class to play.
The changes to wizards are a different matter, but I'm personally from the "classic D&D wizards need to be evened out" school.
In 4e, one of the benefits of broader classes with "sub-builds" is that it opens up space for the sharing of utility powers (and in some cases attack powers as well), which reduces power bloat. 5e seems unlikely to have a 4e-style power system for non-spellcasting classes, which makes me curious about what, if at all, is at stake in this issue.
For example, if (i) the fighter class has 3 build options, and (ii) being a fighter doesn't involve any mechanical choices other than choosing one of those 3 builds, then why call it one class rather than 3?
For clerics and wizards choosing spells it probably will matter, but I worry a bit about a return to pre-4e overpowered, in part because overly broad and flexible, spell casters.
I see two themes from your post here, that I would like write about. The first is the bit about the three fighter builds. I think the example you gave is a good one for the discussion. To me, a fighter is just one who fights. Now with that statement I can make a good deal of character concepts. Now sometimes a concept or a build is good enough to develop it into its own class. For example we have barbarians, paladins, and rangers who also fight and are designed with the idea that they can do it very well (speaking from 3.x/PF frame with regards to FULL BAB progression).
The idea of paladins, and rangers, and barbarians could all be made with the fighter class thematically.In fact, my favorite paladin that I ever played didn't have a single level in the paladin class (his charisma was too low)but he still went the route of the paladin with fighter levels and it worked out just fine. But all three of those classes possess something at their core that the fighter does not as a class. Be it spell casting, animal companions, the ability to rage. Those crucial differences, the bits that cannot be picked up from say a feat, are the things that merit the design of a new class to me. I don't for example think a dexterous fighter who has chosen to focus on thrown weapons and stealth deserves to be a different class from the fighter in armor standing up front take it in the face and dishing it out in return on the dime of her blood.
The idea that I could make many builds and or flavors within the rule space of a single class is something that appeals to me greatly. The idea of making two classes out of variations seems like a greta way to waste pages as well. In essence the fighter is more than just the sum of each specific possible build you can make, it is the collection of ALL of them. When I set down as a player in a group and we are discussing group viability three things usually come up-
How will we stop the bleeding?
How will we keep the bad guys off of the ranged toons?
How are we gonna open that door?
I have found the answer to those questions will vary from group to group (sometimes the second question doesn't even need to be asked). I like that in the D&Dish games I play that I could answer any of those questions with almost any class. That means that I as the player decide the role of my character. I don't have to be shoehorned into anything by mechanical role expectations. I do think that groups of players should sit down during the toon making process and decide how certain eventualities will be handled, and the idea of roles as a concept fits very well there. I just think that the broader classes give the game variety. When you cut anything into small enough pieces it all looks the same.
The second thing I gather from your post is a theme of dislike of casters power (in the Vancian world at least). I have heard this argument before and it has some validity. Full casters can and do break the game in many ways. I tackle the issue by really only letting players in my games have wizards after I have shown them the role that I want them to have in the game world. Its tough when you get a player who has a full blown caster and instead of making their own niche they go after the niches of the other players. But I find this to be more of a player issue than a class issue. But hands down if someone wants to ruin everyone elses fun with a full caster, they can do it by the rules.
I also break the spells up into lists regardless of the source. The lists I keep are high arcana, elder arcana, and banned. These lists started in my groups towards the end of AD&D2e. And the players made them not me. The players put the spells on the lists. And we basically handle them like this: A caster may only cast one spell from the high arcana list in a single encounter for any reason(And they may not cast a second if the effects of a previously cast spell from this list are still active). A caster may only ever have one spell on the elder arcana list memorized at a time, this spell may not be cast during combat, and this spell cannot be dropped voluntarily, if memorized, it must be cast. and a banned spell is simply banned.
I mention these house rules we use because I think that WOTC could really improve the vancian system by leaving it largely open but perhaps adding a similar keyword system to it. It would make it very easy for a DM to say we are in a low magic world mages don't get the high arcana-ish spells (I have done this in the past and my players had a great time). I am not certain what type of keywords would be appropriate for a core adjustment to the game though, I just know what works for me and mine, and it took us years to make our lists (the lists are player maintained in my group, so for a spell to get on them they had to be mentioned by one of the players and discussed by all of them to actually get on the list. I didn't just fill them as the DM, although I did offer a list of spells that I wanted on it, and let the group decide).I also think that organized play could do with a banned list for sure (kind of like MTG) to keep silly combos out of the organized play sessions.
love,
malkav