Since it is D&D, we are going to have classes. As much as I wish we wouldn't, we will. That said, please be aware that what many people see as faults in various rules boil down not to problems with the rules they state, but with poor class design that forces people into incompatibilities between what they want to do and what one of the designers thinks they should be doing.
To sidestep into a brief example to help illustrate that last statement, lets compare PnP roleplaying games to something like a computer game, say the original Neverwinter Nights. While you could move your character around however you wanted, could fight however you wanted, the inherent limitation of the medium meant that you could really only do what the game designers could anticipate you doing. For example, if you had a character with insanely high strength you couldn't just kick through a wall to get where you wanted to do, you had to find the proper door. If the writing for the game as poor, you as the player might be presented with two options to choose from while as a player you immediately see an obvious third option that would solve everything, but the writer didn't think of it, so you can't do it. The advantage to pen and paper roleplaying games is that we have so much more flexibility, that we aren't given as many arbitrary "No, you can't do that because I didn't think you would ever think to try it" moments.
However, class design is one of those areas. Some classes are broadly defined and people generally have no problems with them. Fighter, Rogue, Cleric, the oldschool stuff mostly. These are great classes because they do not carry much if any baggage as to what they should or should not be. The Fighter is just as good with a 2h sword as they are sword and board or with a bow. You could have a big strong tank of a character, or you could have a lightly armored nimble fighter with a rapier. The class is presented as being an empty framework, and that made it strong. Same with the Rogue. It wasn't pidgeonholed as a thief, it was given a broad toolbox and could be a catburgler, a professional dungeoneer, or even a master craftsman. Broad concepts lead to strong class design that doesn't stand in the way of what people want to play, it gives them the tools to make whatever they want.
Poor class design relies on stereotypes and limited scopes. The Paladin falls into this, as does the Barbarian, the Bard, the Monk, pretty much most of the newer classes. Why is this a bad thing? Because to differentiate these classes from the broad class bases (after all, the Barbarian and Monk concepts can both be fully realized with just a Fighter class and some creative descriptions, and the Cleric can already wear full plate and cary a sword and shield, so they already fit the role of the generic holy warrior) they are given unique abilities that generally can't be found anywhere else. However, because the class is based on a narrowly defined stereotype, it comes with restrictions that mean it can only be applied to that stereotype. If you want to use the mechanics they offer, then many times you have no choice but to adopt the stereotype.
Granted, there is a good reason stereotype classes exist. Many players are limited in their ability to make something from scratch. It is easier to take a stereotype, change one or two minor details, and then run with that than it is to take broad concepts and really define everything yourself. That leads the broad classes to look boring because they don't hand you everything on a silver platter ready to play. One of my old sayings was "If you can't make an interesting human fighter, then you aren't ready to play anything else yet" to mock that particular mindset.
It would also greatly decrease, IMO, the noise surrounding other mechanics that work just fine in their own right, but cause problems when combined with restrictive class stereotypes. Alignment is a prime example of this. Dig up any thread about why alignment is horrible and you will invariably find talk of Paladins, Monks, Barbarians, Warlocks, etc. The problem isn't with alignment, it works fine. The problem is that you have given special abilities to one of these stereotype classes that you can't get any other way, and people are railing against the restriction, whatever it may be.
Now granted, a competant DM can and should wave stereotype restrictions when the mechanics are being used in a non-stereotypical way, but to quote the Oberroni Falacy (is that still something people on the boards here recognize?), "Just because Rule 0 can fix it doesn't mean it wasn't a problem to begin with". Using 3e as an example, what if I wanted to play a Dragonball Z style character? We had the rules to cover that, you could replicate just about all of it with a Monk/Warlock. The super powered characters from DBZ were just Gestalt Monk//Warlocks. Oh but wait, both of those were narrowly defined stereotypes with conflicting flavors and unique abilities that couldn't be otherwise replicated. Namely alignment restrictions again. Apparently the ability to punch someone really, really hard is something that only Lawfully minded characters could do, because it was based on the incredibly narrow scope of the old kung fu movie shaolin monks. And the Warlock had no choice but to make deals with devils or chaotic entities to be able to shoot energy out of his hands?
If those stereotype classes had been better designed up front to give us the tools to make what we wanted out of them, and not been so restrictive in what boils down to nothing but one person's idea of appropriate flavor text, then there would have been no concerns, and everything would have run smoothly. Instead, we get issues where people bash alignment as being broken because a designer decided to use what amounts to the same restrictions as PrC for a base class. The "You must play this character in the preset way we said you could" is damaging to the game (and not to start anything, but I think its one of the biggest problems 4e had, so many fewer choices meant much less room for personal design and creativity).
Please, keep class design broad and generic. Make a Rogue, not a catburgler or a pickpocket. Make a Fighter, not a swordsman, a Wizard, not a Pyromancer. Classes should be tools to help you get the mechanics for a character you want to play down in a useable form. Classes should not already be the character and you just pick which color eyes they have. Thats what computer RPGs are for, and this is not a CRPG.
i am against generic classes, i found them to be boring and uninspired, infact that was my problem with most class selection on 3e...so many classes, yet they were just slight variations of each one, very few felt special and inspired from the mechanical point of view (warlock being one of them, thought 4e's warlock blow it up from the water).
If classes are generic and uninspired, you can't make a cool character concept with them, a character is what he can do after all, and if the character can do cool stuff...well, i don't think i need to explain that.
Also a huge chunk of new players would come from computer RPG players (all the D&D players i currently play with, started with computer RPGs)
I'd prefer all classes to be a generic as possible/reasonable. And then let pre-built builds be published (possibly in the same rulebook) and given flavorful names and descriptions to help people select certain styles of characters.
I'd be a fan of broad and generic classes if they were:
1) not prone to be unbalanced messes 2) didn't practically require additional resources and add-ons if the player and DM were not in perfect synchronization.
But they often do.
I don't mind classes being broad but they tend to need attachment to a concrete archetype. Not overly focused but somewhat.
Orzel, Halfelven son of Zel, Mystic Ranger, Bane to Dragons, Death to Undeath, Killer of Abyssals, King of the Wilds.
The class abilities are generic enough, and have been since Fighting Man, Magic-User, and Cleric were little more than fantasy-repaint analogues of infantry, artillery, and support.
The real problems lie in baked-inflavor-mechanics, such as alignment locks, race-locks, Druicide, and "you get your awesome powers from a deity/demon/bloodline/bat-poo/whatever/etc." "This class could represent..." is always preferable to "this class must always be..."
I'm not a huge fan of a million different character classes. It is important to me that there not be many different ways of expressing the same character, because the character herself is the truth and the rules are only a lens for viewing that truth - it cannot change who that person is, or what she can do, based on how she is viewed. There should be no question over whether your character is a fighter, wizard, rogue, or cleric; at most, the question should be of what degree an individual is best represented through direct combat or subterfuge, arcane or divine magic.
(To be honest, I'm not even really happy with rogue being its own class here, unless "non-combat-guy" is a serious role that people want to play. I would be much happier switching the fourth class to "archer" so everyone fits on a neat two axis system between melee and ranged, black magic and white magic. That way, class only governs your role within combat, and we can leave skills and non-combat stuff to draw on entirely separate resources.)
From my perspective, it seems that many classes exist only to address shortcomings in the base system. Monks exist because they wanted unarmed/unarmored combat to have a place within the world, but the system rules clearly state that any weapon deals a lot more damage than no weapon and that more armor is always better. If the base rules had been less restrictive (maybe everyone uses Wisdom to AC while unarmored - even wizards - and maybe unarmed damage could increase along with your BAB by default, just as an example), then we wouldn't need an entire class just to cover one niche fighter.
The 3.5 samurai class exists because they wanted it to be possible to dual-wield heavy weapons (well, one heavy and one light - the goal here was to let you use Strength to hit and damage) while wearing heavy armor, but they'd already clearly written the dual-wielding rules to make that nigh impossible (requiring ridiculous Dex investment for the two-weapon feats, while preventing you from using weapon finesse or gaining the Dex to AC that you would get out of light armor).
Aside from obvious ambiguity in representation, this is also an example of the dreaded trap option. Imagine this exchange between a new player and a DM: : "I'm going to make this awesome samurai lady, and she's going to be so cool with her naginata and light armor." : "Okay, but just be sure to avoid the samurai class, and look into taking barbarian levels if you can."
I'm not blaming anyone for the samurai class being less powerful than maybe it should have been, I'm just saying that increasing the number of classes reduces the consistency in quality of each class. If they make a ton of classes, then some of them are going to be more powerful than others within their area of overlap. Reducing the number of classes aids in this.
The interesting part is that, between themes (or whatever they're calling feat-chains this month) and backgrounds for skills, it almost looks to me that they're actually approaching this direction for 5E. If you want to fight head-on, then you're a fighter, and if you'd rather sneak attack then you're a rogue. If you want to cast divine magic, then you're a cleric, and if you want to cast arcane spells then you're a wizard. Want to play an unarmed fighter? Decide on a direct or indirect approach, and staple the brawler theme onto your fighter or rogue. Want to raise undead to fight for you? Decide whether you do it for yourself or on behalf of a higher power, and then staple the necromancer theme to your cleric or wizard.
Right now, the only stumbling block I see is the alternate-method spellcasters. The line between wizard and sorcerer has always seemed a little blurry, but here it would present a direct obstacle to the ideal of obvious and incontrovertible class distinction.
Right now, the only stumbling block I see is the alternate-method spellcasters. The line between wizard and sorcerer has always seemed a little blurry, but here it would present a direct obstacle to the ideal of obvious and incontrovertible class distinction.
Sorcerer is a stock wizard with the spellcasting system swapped out, and little else.
I am hopeful that backgrounds and theme/specialties are beefy enough to create unique and effective characters without being strapped with the baggage associated with a pre-made class. But some players don't mind using them and that is okay, too.
D&D is overall a class, not skill based system, that has been what at least up until 4e made it different from other competing systems.
Why would you want to make D&D something that it isn't, rather than playing something that fists your need? Lots of other games have 'broad' classes.
Take the cleric for instance, who up until 4e, could turn undead, and ONLY the cleric could turn undead. That was a skill in the domain of the cleric. Now people want that to not be the case.
If D&D didn't have its archetypes and locks, it would just be another generic RPG is the sea of them that exist. Its class-based play makes it truly unique against its peers.
Take the cleric for instance, who up until 4e, could turn undead, and ONLY the cleric could turn undead.
That's what a Class system is for.
If I want [ability A], I'll pick a class that gets ability A in its class abilities. If I want [ability A], I do not want to be stuck with the fluff insisting on who my character is and how he's going to behave.
Any wizard can have mall-hair and claim to be from a long line of spellcasters or a mutant or something. Why is Sorcerer obligated to do so?