I am wondering about everyone's thoughts on the age-old debate of how character development and advancement should be addressed.

History has shown us that RPG's tend to fall into three main categories of character development and advancement.

1)  Caste/Class/Profession:  A character chooses a profession and the profession (oftimes called a class) dictates the abilities, skills, and powers, as well as the ability to perform such things. (D&D first and second edition, several others)

2) SKILLS: There is no class, only a set of skills to choose from, and how you spend points, or slots, or whatever dictates your degree of proficiency on those (GURPS, Fantasy Roleplaying, many others)

3) Melange: Players choose a class or profession and are given a set packet of abilities (or skills) and then get to choose some to personalize their character.  (D&D 3rd, RIFTS, many others)

If find that each system has advantages and disadvantages.  Caste based systems tend to be very limiting and every character of a given class becomes an archetype, easiliy identifiable and just another flavor of vanilla.  Skill based systems tend to run into "nada" syndrome which means that one inevitably finds that they cannot take a crap because they didn't take the "potty" skill.  The melange system tends to harbor the worst apsects of both, you not only are like everyone else, but you also cannot take a crap.

The advantages are, of course, legion; and fo rthe sake of brevity I shan't go into them.

What type iof charcater development system do you prefer to integrate into your games and why?

All three. At least thats what im trying to work on.

I like the idea of not so much a class, but a general Idea of what that character your going to create is. An Archetype, so to speak. In the bases of terms I'd like to define that as DPS/Tank/Healer/Support. (there could be more but Id like to stick to a few groups.) Sort of what 4e is doing, they are making archetypes, and this even goes as far as over to the monsters side, with mooks and grunts and bosses. But as you point out it seems like another flavor of vanilla. (Oh hes playing another fighter ). But id like to take the class system and elevate it a bit further. Note that 4e is trying to do the same with their version that a fighter can be this or that (Sword and board, or etc) depending on their fighting style, so that a fighter, just insnt a fighter anymore it could be x or y type of fighter. Or that a wizard is more specalized here or there. (maybe I should stick with 4e huh? )  

Next, skills. Id like my skills to be a bit more complex and simple at the same time. Meaning Id like that a player would be able to start out with a array of skills. Now, depending on the archetype they picked they might get some other skills or bonuses that some on outside their area couldn't get right away. But Id like to see simple skills or base skill sbe able to evolve over time into something talored to the players needs/wants. Example Cooking. Every one can throw something into a pan over fire. That's the basics of cooking. How well you can master this will determine how well the food is cooked. However there are some considerations to ponder. At a certain point Id like the player who has cook with enough use to "level" it up, now not only do you have cook (untrained skill or mundane) but you get a skilled application (Preperation) so now you not only know how to cook, but you know how to prepair it right and what seasoning to use, and how long to cook it. Then as you use thouse skills you can further "level" it up into a master skill, and it would get the aspect "Presentation" so then you would have a skill that would allow you to preperate and presentate food that you cooked to perfection.

anyways im rambling but im still on the fence deciding about alot of things atm. =(

I prefer what I guess would fall under melange, or I'd call it a hybrid.  Those melange systems you mention play very different, some are more on the class/caste side, others more on the skill side.

Here's what I like best:

A certain constellation of skills at a certain level can define a caste/class/etc, but in the sense that if you make the requisites for that caste/class you get added advancement benefits.

In one implementation, you slot into the caste/class benifit track you can access early on.  There is nothing to prevent you from obtianing skills outside the constellation of that track.  In addition, the benefit constellation is fairly broad so you have some chose in the constellation of skills that qualify you for the added benefits.

Ease of skill improvement is based on character traits, "talents" for learnig certian kinds of skills.  In a sense, a given set of talent scores could define a castes/class etc.

In another implementation, talent scores are bought initially.  Added advancement benefits are accessed again by achieving a certain constellation of skills at a certain level BUT the benefits are organized by the talents instead of some predefined caste/class set of benefits.  For example, raise enough combat skills to a certain level and get added combat oriented benefits.


ADVANTAGES/DISADVANTAGES

I see them a bit differently.  

Caste/class in the extreme case can be very limiting with narrowly defined sets of skills and suffer from the "nadda" syndrome if your class does not explicitly have a certain skill.  But in reasoned implementation, i.e., easy to design around, a certain base skill is assumed in ordinary things.   Another extreme is class proliferation where you end up having more classes than all publised skill based games combined have skills.  At it's best, class based systems provide focus, balance and ease of character design, with some flexibility in "skill sets" within the class.  

The big problem with class/caste is you lock into a pre-defined view of constellations of skills which may not mesh well with genre.  Nevertheless, the caste/class is the easiest to implement IMO.

For Skill only focus, the "nadda" syndrome I see only as an extreme of over zealous application and not a valid problem of the approach.  That is, it is easy to design around, simply don't require skill checks for everything (especially mundane crap) and add some default success.  The bigger problem with Skill only is one of balance, can someone just focus all their improvement into one skill in a way that destroys the gameplay you want. What are the limits on which skills you can improve and how fast?  If there is a small dynamic range or slow improvement rate then not such a big deal.  


I see the Melange as the preferred approach but it requires attention to the details.  I don't see how it is as limiting as the class/caste as it has built in ways to differentiate (that is why it is melange and not caste) nor as susceptible to the "nadda" syndrome (which I think is a more common failing of class/caste) because it makes use of class/caste to provide broad definition and skills for details variation.  That is, there is not as much the danger, as in SKill only, of every little thing having a seperate skill.

My favorite method isn't in there anywhere.

-clash

A well designed Melange system is my preference.  It gives you a decent starting point that you can work off of but still the freedom to develop your character as you see fit.  The key being that you are allowed to deviate from your starting 'class' and change roles as you further your goals or your characters interests change.

There isn't the one-size-fits-all solution that is superior in every regard.

In my current system Mazeprowl, I used a melange solution. There 24 skills, divided among the 4 archetypes. you choose a main archetype (improves starting skill points for the 6 skills of that archetype and makes them cheaper to improve) and a minus archetype (decreases starting skill points for those 6 skills and makes them harder to improve).
Advancement is ye olde "spend your XP as you want".
Therefor I get a flexible improvement system, but the archetypes protect the niches.

For character creation, instead of the proposed solutions I use a priority grid as in TRoS or Shadowrun 1-3.
There are 5 priorities:
- 1x Attributes
- 4x Skills & Starting Equipment for each Archetype field
It gives some strategical choices from the start off and forces you to spread your skill points a bit instead of getting a one-trick-pony with maxed out gun skills that can't drive a car, but still protects the niches.

A "melange" or "caste" (nice expression) is OK as long as the system is playable without it.
It's a weak system if a party must consist of "caste"-members, or even "melange" characters. If a skill based member is as producive as the rest, then it's a-OK.