Now, which one would be better to roleplay in a D&D game?
I've seen good characters following either model, so I will say that neither is inherently "better".
It's going to take more contrivance to make a skill like Profession: Cook as useful as Sense Motive for a detective.
So our Detective has only one skill point, and must select either Profession: Cook or Sense Motive? Frankly, if some skills are useless, then I consider that poor game design. If being a great cook will have no in-game benefits, then the player should be allowed to define his character as a great cook with no cost in character resources. Can we ever ensure each skill is equally relevant and influential? Probably not. But we can certainly work to ensure that players are not required to invest character resources into abilities that will have no value in the campaign.
Feats are rather significant investments for people who are not fighters.
Absolutely. So the investment of a feat should have benefits in-game. One such feat is Skill Focus. If it will have no in-game impact, perhaps it is again appropriate for the GM to mandate that character resources are not required to be expended for the character to have that ability.
Ah, so you'd give Skill Focus: Craft: Poetry additional uses.
I would more likely tell the player that "Craft: Poetry" is not a skill in my game world, and that the ability to compose poetry is a part of Perform: Poetry Recitation. Perform, of course, being a very important skill to a Bard. Or I would simply say "Your ability to compose poetry will have no game impact. You can define your character as a skilled composer of poetry without spending any skill points or feats."
Hold on. Are you suggesting that practice combat plays no part in how a character gains experience? Then what are fighting academies and martial arts classes teaching in the in-game world?
By RAW, practice combat plays no part in how a character gains experience. How is it that you consider the inhabitants of this world will notice and quantify the various results of leveling up, enhancing save bonuses or BAB, etc. but they will not notice that activities which do not generate xp do not result in any improvements, however long they are undertaken?
And you ignored completely the part where I suggested that they'd be sent out on missions. Which is another way of saying they get into combat.
So training in any field - cooking, practicing law, playing the flute, weaving cloth, any skill whatsoever - will see the student sent out on dangerous missions involving life or death combat? Regardless, by RAW, the experience gained will not vary with the skill of the teacher, but the number and nature of missions he sends his students on.
If we accept experience as a game mechanic only, simulating the many ways in which the PC's could be improving and enhancing their skills, then there becomes an in-game reason for such training. However, the acceptance of game mechanics as simplified abstractions would also seem to eliminate the ability to assert the game world residents can quantify those mechanics, since they are only abstractions of the game world's reality.
The mechanics of saves are quantized; you can fail 10% of the time, you can fail 15% of the time, but you never get someone who fails 13.5% of the time. If someone does a long study of these probabilities, they might notice it. Historically, this sort of thing has happened; during World War II, South African mathematician John Kerrich carried out 10,000 coin tosses while interned in a German prison camp. (He must have been really bored.)
So some Wizard will hire someone to stand and have Sleep cast upon him 10,000 times to record how often he falls asleep? Presumably, he must hire many people of many different levels of experience, since he needs to quantify when they obtain a bonus that increases the probability of success. He also needs to find a reliable measure for their basic statistics as WIS bonuses can throw off his curve.
I suggest, rather, that the "increments of 5%" save bonus are an abstraction necessitated to make the game playable, and are not observable or measurable within the game milieu. The characters might well reason that sometimes the effect of the spell is resisted, and even that persons with a strong will are more able to resist the effects of those spells. Demarcating what causes such enhancements to one's will, and in specific demarcations? Not so much.
How does it enhance the character more to take more levels of Oracle? Is it conceivable that one level of Oracle might be enough for a character concept?
Quite. What I consider not to be conceivable is that each and every Barbarian will attain precisely one level as an Oracle touched by divine forces, with each and every one of them rendered lame as a consequence.
Do you remember the parts of 3.5e rulebooks where it discusses how you can adapt classes and prestige classes? For example, Champion of Corellon has an adaptation section where it talks about changing the deity to another.
Do people not do that in Pathfinder? Are the class descriptions straight jackets now?
To me, those adaptations are intended to create interesting aspects of the campaign world, not to provide carte blanche for individual customization. But then, if I want absolute customization, I would look to the Hero System where all abilities are open and available, within the constraints of available character points and any restrictions imposed by the specific game.
Hold on. Nothing in the Challenge description requires you to warn enemies of your tactics. You simply single him out for more damage and make yourself easier to hit by other enemies. You do not have to be honorable to do this. You could be a some sort of psychopath and still focus your wrath on a single foe in combat.
"Once per day, a cavalier can challenge a foe to combat. As a swift action, the cavalier chooses one target within sight to challenge. The cavalier’s melee attacks deal extra damage whenever the attacks are made against the target of his challenge. "
This seems to imply a positive action on the Cavalier's part - throwing down a challenge, not just picking a target. How is that foe "challenged to combat" if not by a challenge issued to that foe?
The Ronin option does not require someone to act honorably - to be kind, just, fair, truthful, or ethical.
These are not perfect synonyms for the term "honorable".
The description of Ronin reads: "Known as ronin, these samurai wander the lands, serving their own code of ideals." He could behave like a Chaotic Evil scumbag - and indeed, the lack of alignment restrictions allows for just that kind of character.
He still needs that code of ideals, whatever it may be. The consistent adherence to such a code strikes me as more Lawful than Chaotic, but it would still only be one aspect of the character's overall persona.
That because Honorable Stand has honor in the title, it requires the character to act honorably, as if there is some sort of truth in advertising policy? Nothing in the actual ability requires you to be an honorable, it just requires you to stand your ground to the bitter end.
Apparently, Samurai consider death before retreat a facet of honor. If he acts dishonorably (retreats from that foe), he loses this ability for 24 hours.
And indeed, Black Jack does exactly that. Tell me what's wrong with it.
And again, I am not saying that every multiclass, or even every dip, is bad. I am saying that simply choosing options with no basis other than "I like these mechanics and they make me powerful" is not, in my eyes, the mark of a good character or a good game.
To the specific character, the only reason for that dip which I see from your comments is "I gain a mechanical advantage". I see nothing that makes him a more interesting character as a result, just "I had one more level and this dip provided a mechanical advantage".
Then how does your in-game world treat spells, abilities, powers, and feats that directly tell a character metagame knowledge?
This would vary depending on the specific ability. It may be an assessment of combat skill, magical knowledge etc. depending on the specific ability, and provides information in that regard in the terms of reference familiar to the character (a mighty warrior; a skilled novice). Again, we abstract that in-game information by providing information the player can evaluate - the mechanical abstractions.