Worth noting:
* In Morrowind, more expensive spells are harder to cast; assuming reasonable attributes and skills, you will notice that the success rate of spells decreases as the cost increases.
* In Daggerfall and Oblivion (as well as Arena), more skilled casters need less Magicka to cast the spell. (One complaint; the difference feels too small until high skill levels, where one point of skill makes a huge impact; a formula more like Arena's, where the equivalent of skill is in the denomination, would feel better here.) Furthermore, in Oblivion, you can't cast more expensive skills until your base skill reaches a specific multiple of 25.
* You can cast cheaper spells more times before needing to rest, hence why leveling strategies involve casting the cheapest spell you can get over and over again. It would make more sense for more expensive spells to be better for practice.

(Side note: I'm thinking of making a SaGa-like game, but I still don't know whether the game will have MP or whether, instead, the limiting factor will be item durability or skill uses.)

By the way, Final Fantasy 2 does have MP that can grow independently of magic skills. The problem is that you need to use (or, more precisely, lose; this matters when MP damage/drain effects come into play, like if you have Minwu cast Sap on the party to trigger MP gains) a significant portion of your MP to get a max MP increase. It would make more sense for the game to keep a running total of spent MP and be more likely to give you the increase when that total gets high enough. (HP doesn't have that issue because it's lost more easily and is more easily restored.)

Side note: In Final Fantasy 5, which I have played and is one of my favorite RPGs, you actually *can* do this; equipping a magic skill on a Monk will actually give them decent magic power and MP.

(Note that I prefer Blue Magic in this case, as there are some spells that are HP dependent (White Wind comes to mind), and Monks have the highest Stamina, and hence the highest HP, of any job.)

Also, sometimes unusual set-ups have surprising usage. FF5 has some rather interesting ones, many involving abilities interacting in interesting ways, particularly with the way abilities grant stats to the user. For example, Blue Mages can use swords (including the Rune Edge, which uses both Strength and Magic Power to determine damage), but have low Strength; add 2-Handed, and not only do you get double damage, you also get decent strength. Said Blue Mage can do a *lot* of damage with that Rune Edge; just put them in the front line with that set-up (but note that Blue Mages don't have the best HP, but at least it's not the worst), make sure to keep MP up (or else the Rune Edge will lose the Magic Power component of its damage), and just swing away!

(There's also tricks like using Goblin Punch with Excalipoor equipped.)

But I *like* teaching characters new tricks late in the game.

Also, in Final Fantasy 5, changing a character's job will instantly make the character decent in the job's main role; a character who becomes a White Mage can immediately use all the White Magic you've collected, and will have decent Magic Power for the Cure spells (and Holy). This is especially helpful if the character you've been training as a White Mage is absent from your party for plot reasons. (There is one part of the game where that's actually quite likely for a typical casual player.)

Realism is over-rated. A game should be playable and fun to play; those aspects are far more important than whether the game is realistic or not.

Having to manually feed your characters (like in Ultima 7) may be realistic, but it's not fun, and it really doesn't add anything to the game (unless the game has a strong survival focus, which Ultima 7 does not). Similarly, having weapons decay over time if not repaired, like in TES games, may be realistic but it doesn't really add anything to the game. (This works better in SaGa and Fire Emblem, where durability limits are clearly presented to the player and feel more like an integral part of the game design.)

Remember: Super Mario Bros. has platforms and ? blocks hovering in mid-air, and I don't see anybody making the argument that the game isn't realistic.

You're talking about a significant amount of simulation to be done in the game. This may not be the best choice for a few reasons:
* It's extra work to implement this.
* There's a significantly greater bug surface. Your simulation could very well have bugs, and the wrong bug could make the game rather un-fun, like having too many enemies or the occasional enemy that's far stronger than what would be considered reasonable.
* I personally don't like having the world change just because I decided to take time to do something else.

In the game I'm thinking of making, the world will either:
* Be static (most likely case, because it's less work)
* Change only with scripted events
* Be treated like Zelda: Majora's Mask; scripted events happen with the passage of time, and the world will end if you take too long, but there's a way to reset the clock and the world.

Then how am I going to get dragon meat for my monsters? I would really like my monsters to get some breath attacks, and possibly even become dragons themselves, so how should I go about doing this?

(I don't like it when abilities like breath weapons are enemy exclusive; in the game I'm thinking of making, the player *will* be able to get access to breath attacks. After all, you can get them in some SaGa games (1-3 and Frontier, though Frontier is the only one where it's really obvious, also a character in RS:MS gets some I believe), and in some Dragon Quest games (5-7 all give you a way to get a controllable character with a breath attack).

I think things like levels can represent threshholds. So, for my game idea, for example, I want to have skills that have levels, but not an over-arching character level. That way, a skill's level can represent a skill's efficiency (i'm thinking about aiming for a percentage, but i could change it later since i'm still in the planning stages). For example, I might have a level for a "jab" skill. Max might be 100, which would be the perfect jab, while it might start at something like 5 at aquisition (0 prior), meaning you punch like a baby 'cause you've done nothing to hone the skill you've learned. However damage would come from strength and perhaps weight (normally you don't throw your weight behind a jab, but it can be done). Strength would be increased by doing actions that use it, including maybe carrying heavy gear (i'm also considering having specific strength attributes per limb, since some skills already will require limb tracking). For games that aren't going as specific as mine, you could still have a similar approach. Also, a "level" could be a culumnation of gains for which a system like "leveling enemies" can occur (skyrim and oblivion both do this, but if you divorced the stat gains from the levels you could have a more fluid system). For The Quest, specifically, it's due to slow you the hell down. The guy who coded the thing left the company, and it uses signed words (15 bit limitation, plus a sign bit) for damage calculations. Remember, the original target was Palm OS, and porting the scripting engine to the PC was supposed to be a quick transition. I've personally talked to the dev about this issue, and he seems to be of the opinion that the current engine is too complicated for him to try to delve into, when he did little to no coding on it inthe first place. He's been busy writing a whole new engine from scratch for The Quest 2, where i hope he's learned from his predecessor's mistakes. The woman behind Zarista games is, well... Let's just say i disabled my account on her forums intentionally 'cause she's, well, something else, and I think her trying to do her porting has caused some issues with him being busy trying to also help her get things working, 'cause, well, she has no clue what she's doing when it comes to scripting and stuff. I tried to get basic info out of her just trying to see what the issue with overleveling was and she said something along the lines of "the engine doesn't like big numbers, i don't really understand it, but i was told i need to bring the numbers down." I've had some problems with her mods, but i just gave up on trying to talk to her, because, despite her admitted lack of knowledge, she tries to apply her authority. But i'm getting way, way off topic here. Which reminds me, i need to convert my android mods to work with the GOG version of the game since i grabbed 2 more of her expansions recently. You would be correct. I'm not sure which box it's in, but I still have the thing laying around here, somewhere. Back then, it always seemed cool to have the ability to morph into enemies and use their skills against them. That was always something i liked about games like that: when you could polymorph and see for yourself how unfair things really are. It's something i'm naturally going to include in my game as well, with the exception of specific NPCs (you might be able to polymorph into a vampire, but not Carmilla, althoguh getting her stats and skills and being a vampire would be entirely possible).

An alternative explanation might be that it's easier to "fumble the mana" but not necessarily harder to cast. I think what they were trying to do was not base skill gains on mana depletion. Consider, too, how mana restoration and fortification would suddely be equivalent to fortify skill potions, at least in lore. I do agree that something has to be done about that simple spell cast to level up strategy, though. And just about everyone i've talked to says something is seriously wrong with the restoration leveling in every game. I try to make decisions based on how i would build the lore, as well. There's no reason you can't do both, like games like nethack and TES do. A wand/enchanted item might have an X number of charges, which adds a dynamic to magic oriented characters by allowing them to pre-charge items to be used themselves or other characters, at the cost of the weight of holding them. I always thought this was a nice mechanic of MMOs like Runescape and Tibia (moreso tibia cause you could fall back on mana withotu runes): you can certainly boost teams this way. This will likely be the approach I take, as well: then I can have NPCs who buy and/or sell charged runes as a "crafting skill" for mages specifically. You would actually likely expect this sort of thing to be far, far more common than what we see in alot of these games. I still think a smaller drain should result better than a full drain (or rather, have a exercise vs abuse threshhold somewhere towards the middle like you should for strength to simulate the difference between micro-tears and full muscle tears). The question is, lore wise, do we want to consider mana draining something that can damage the brain? If you have trouble lifting a 50 pound axe, and it drains 50% of your stamina in 1 swting, you should probably get injured for that, pulling a muscle and lowering your strength. Sometimes it does make sense, depending on the type of monk. FF has been inconsistent on this matter. Is a Monk someone who has magic powers via religiosity or is it just a dude studying martial arts? It's kinda like how "paladins" in some games are just generic fighters, but in others they might be white mages who can also smack pretty hard with a mace. Front line glass cannon. I do like it when you can create your own classes, don't get me wrong, but it should make sense. For example, the "religious monk" might just be a healer or an offensive mage that relies on fists or daggers (i still say daggers are a martial arts weapon, not a knight's weapon) as a backup weapon. Note though that your blue mage quickly ends up a one trick pony that's in deep trouble if anything smacks it good enough. There should always be a tradeoff. I think that's a separate issue, though. I think alot of games lack in the amount of tricks and skills (largely due to having to animate them), so something like your monk developing a new technique rather than just getting stat boosts is unheard of. Meanwhile, this is something i hope to change with my game as well. Since each normal human turn will actually be several internal turns, inventing your own skills from scratch as a player will be entirely possible. In real life, we have the art of "Iaidou" in japanese, which applies primarily to their long swords. I, taking some of the same principles, applied it to my folding knives: I can open a knife and attack in one motion. I want players to be able to come up with this as well: so they can learn skills from books or something that they might not actually apply directly, but turn around and re-apply it to another skill. Maybe I have a boxing character that knows punches: why can't we just make a new skill that applies the same techniques but with a knife in the hand (which happens in philipino martial arts)? Something like a hook shot with a knife might start at level 90 if the empty-handed hook is at level 100. Yeah, i'm not particularly fond of that, though. Sure, it might be convenient, but, then again, I hate character separation for plot. If my characters are going to split, they're not really under my control if they just do some pre-coded split (unless we got arrested or something like that). This is always the challenge of plot centered games: sometimes you want certain things to happpen a certain way for your plot, but maybe your plot isn't particularly realistic if the player wouldn't naturally go along with the plan you had laid out. I was thinking about, for plot reasons, making a bunch of master vampire clans, for a starting plot. Each will represent some sin, but will be friendly (other vampires lines might exist but aren't plot related), and you can either join their clan, reject their offer and leave peacefully, or you can choose to "end them." Depending on your decisions, the other vampires might act accordingly, as well as the non-vampires under their rule. Do you let Carmilla, for example, rule as a misandrist, because of her lust for women and her superiority complex as a master vampire and a dictator, or do you prove that men are worth something (if you're a man), or do you free the enslaved men by force? Maybe there are some other angles you can take, including ones i never thought of, simply by modifying your relations with her to the degree that somehow you take power and can order her around if you end up a stronger vampire and she recognises it.

Well, yes. In terms of gameplay mechanics, that power trip is what I play for. Definitely if it's a single character game, which I also prefer. Restricted classes, while always remaining confining, may work for games with parties of 6+ (smaller parties usually mean you can't get all skills if class-restricted), but otherwise it's just frustrating.
Don't play for the challenge. And I'd hope there had been plenty to do until that point and may still be later, experiencing the story, the atmosphere, character interactions, exploration. And, yes, also doing what's right in that world.

I think RPGs should make you define your character through some limitation. A lot of JRPGs and Diablo style games do this by making you pick a class with limitations based on that class, which I think is kinda lame personally, but in that kinda system I don't think a cap makes sense. You're already limited by the class, so why not just let you get more and more powerful.

A lot of Western RPGs however let you define your character by choices you make, such as which stat or perk to add to when leveling up. In these sorts of systems (Fallout, Cyberpunk, Gothic, whatever) I think a cap is important because it prevents you from becoming a master of everything. In Cyberpunk for example there are 5 stats with 20 levels a piece, but you get like 60 points the whole game. So you can max three stats out of five, with the other two getting almost nothing, or you can max one stat and spread the others out. It lets you define your role and style.

I like this kinda thing with factions and stuff too, like if you create a melee warrior with no magic skill I don't think you should be able to do mage quests. That's a big thing that annoys me in Elder Scrolls games. Save that stuff for a future mage playthrough.

That's actually an issue in Dungeons and Dragons as well. If you try to scale the system up to extremes, you get issues like the following:
* If playing with physical dice, how are you going to roll 305d6 damage? (There's actually an epic spell in the 3e Epic Level Handbook that does this much damage; it's a difficult spell, and I actually think it's more difficult than it should be (because it uses an overpriced seed (Destroy); there's underpriced seeds like Fortify which is how you break the game with epic spellcasting.)
* The mechanics heavily use d20 rolls that add values based off those numbers. When the numbers get big enough, they overwhelm the dice roll and a small difference has too much of an effect. Around level 4000, for example, a 1% level difference can mean the difference between a hit chance of 5% (minimum) or 95% (maximum); that's ridiculous. (Then again, nobody would play characters of this level; imagine having to go through character creation, choosing one feat per 3 levels, which is over 1,300 feats at this level!)

If you want a game that handles large numbers better, the Disgaea series would be a better choice, and even then you see issues. (Disgaea 3 has an issue where, eventually, *all* attacks are one hit kills; at this point, HP, DEF, RES, and healing becomes useless, significantly reducing the strategic space.)

Right, there's a time and place for everything. My reason for making this game is to address shortcommings i'm seeing in other games. There's way too many hard-line restrictions and not enough organic restrictions. Notice your examples, you have counter examples. For mechanics like weapon decay, it should matter what kind of target you're hitting. Are you hitting a really soft target (like butter?) or are you hitting a really hard target (like stone)? A realistic way of handling weapon decay, especially with realistic parameters, could be made in a way that's still fun. If i have a pretty sharp sword with a solid material, and my opponent is defending with his axe, i should be able to break his axe in a few swings if it's got a wooden handle, and vice versa. Now if i'm trying to beat a slime with said sword, it should really do little damage to the sword or the slime.

The trick is, is it, as you said, actually integral or is it tacked on as a gimmick? And, is it intuitive? People don't sit there playing morrowind thinking about when their sword is going to break, and then it does. Naturally, this doesn't feel right at all. If your sword is being used to hit alot of hard targets, maybe it should slowly get duller and duller, prompting you to sharpen it, rather than completely breaking on you without warning. Maybe if the handle's getting loose from all the hard object beating, maybe it should make some noise or feel wobbly or something to properly warn that something's wrong.

Maybe the problem isn't the realism, but the attempt to be realistic without actually being realistic. The player should be faulted for their own choices, and also gain benefits from making smart choices. Some people certainly do want more, while other people are more interested in the simple mario. There seems to be a pattern, though: people want something that's simple to get into, but has enough complexity to reward the investment into the deeper side of things. Most "mediocre" games fail in the latter, but then you have games like Nethack or fighting games that fail on the former. I think this is over-estimated. I'm specifically choosing my interface for this reason. I had all kinds of ideas for games i could make that solve shortcommings: X and Minecraft hybrid, a better Terraria, and this idea. It comes down to what is reasonable for the interface. For example, my game plans on simulating romance, sex, etc. Obviously if i'm doing 2d or 3d graphics, this is going to be way, way too much work. However, since i'm using text for a tile-like interface, this isn't actually unreasonable. Moreover, intelligent creatures will all be based on one another. So a bandit will contain the same villager AI with some minor changes, like having a separate faction. Same with a vampire: it's just a villager that hates the sun and has a need for blood, and maybe has some special innate skills. A mummy is a zombie with a wrapping. Sure, implementing the villager will be alot of work, but on the flip side, I'm going to be recycling alot. I'm building a game i want to play, not one i plan on selling. I plan on making it such that there are always reasonable steps to take in every situation. Nethack is a classic example in this regard. It always seems hard, unfair, and unreasonable, but someone who has played it for a long time will tell you that, if you're relying on the RNG to give you a good result to save you from death, you're not playing smart at all. Meanwhile, the RNG is often more reliable than it always seems to be: mathematically speaking, if you eat something that is only dangerous to eat 10% of the time, it's going to catch up with you, and you'll hit that 10%. If you're seeing scary things, you should not be engaging, and can often worry about it later. This also means that i'm going to get alot of play-testing prior to release (which allows me to play with the ranges). I'm trying to build a simulation and eco-system that more or less plays itself, and you're merely just another part of it messing with it. Look at the success minecraft has had with this mentality. Yet the overwhelming complaint i hear all the time from people is that "my choices in the game are just shallow and don't really make a difference." I certainly understand the feeling. Obviously you won't be my target audience, if you're afraid of harsh consequences. However, the bandit's not going to get very far in levels if it kills off the nearby village. Meanwhile, if villagers are getting killed all the time, they might try to go search for the camp, or lay traps for the bandits that get them killed. If you miss out on the reward for bringing the head of the bandit, that's your own problem. The village will likely try to take care of itself, as will the bandit camp if you find a way to join them. I'm certainly not your target audience. The thing i hate the most is a doomsday counter in a game. Perhaps, though, we're looking at different levels of importance for characters, as well. My vision is that a player can have any level of importance on the world that they want. If they just want to live life as a slave, that's their choice. If they want to be a hermit that's their choice. Maybe they're the blacksmith that makes equipment for adventurers that might try to try their luck slaying the nearby dragon. Or, maybe you just want to kill the king and take his place. If the world ends, it's because of you, not because you didn't do something. Maybe, though, you just make the world a better place to live in: you establish the world's first cities protected by an armed guard, a nice defensive wall you built, with a library of skill books to be proud of. Maybe you take that city and make it full of just vampires. Maybe it's just demons. Maybe it's all from a certain religious order. Maybe in your city you have peace between the vampires and the religious order. Maybe you take your city and turn it into a massive empire which can rule with an iron fist or exists solely to topple tyrants. Your personal alliances will affect how those people act (for example, if you befriend mostly peaceful people, natural selection will keep things peaceful if you are a ruler).

You, on the other hand, have a "chosen one" design. Of course, because there are cataclysmic events, they must be hard-scripted, and there must be some compulsion to keep the player on the rails, even if it's "open world." Because of the cataclysmic nature of the events, player choice is fairly limited, because there's only so many outcomes to these fairly limited and dramatic events. That's fine, too, but I think that market's pretty well saturated already. If you really wanted to, i'm sure you could seek them out. Eventually, you might find that dragons really like high places. If you want to take on the dragon, as a food source, you want to make sure that the dragon camp has managed to grow enough, otherwise you're not likely to get much dragon meat before making them extinct for the X number of chunks. As the dragons will likely be intelligent, but non-mamillian, maybe you might be able to barter with a dragon camp if it seems friendly, for alternative ways of accomplishing. Perhaps they can teach you certain skills in exchange for providing them with meat instead of them having to come down and hunt for themselves. If they're hostile, maybe you can earn enough money to buy a small team of mercenaries that will help you take on a dragon camp, or maybe it's only 1 dragon.

But, I was thinking that fire breathing should be limited to certain species. If you want tame dragons, you could polymorph allies/pets, you could try to recruit dragons from more friendly means, etc. Odds are, though, that you're not going to just charge at a dragon with a fresh character. They're big creatures, with strong abilities, and they might be big enough to swallow you in a single bite. Obviously you'll need some armor that can resist what ever breath skills it has, and you'll probably want to cast magic or shoot arrows from afar rather than try to take on the great strenth of the dragon up close. Good enough armor will protect you from it pouncing on you or killing you in a single swipe of it's claws.

I'm thinking along the lines of:
* *Everything* is an item with limited durability, or a special ability with a use count.
* There is no MP; the only thing keeping you from, say, casting Flare all the time, is that the spellbook has only 10 uses, and when it runs out, you need to get another. (Example taken from SaGa 2.)

With that said, I do like the idea of creating and charging your own magic items. I remember Dungeon Master, where you could turn an empty flask into a potion with a spell. There's no healing spell, but there is a spell to create a healing potion, and the game tells you the spell early on.

Well, it's a good thing I have someone who can summon Phoenix!

(And, if you're worried about the MP cost of that summon, there's a Blue Mage that can restore HP/MP at the cost of the caster's life, and again, that drawback is easily fixed with another Phoenix summon. In the meantime, note that Phoenix *also* restores MP when it revives.)

Or, if I have a good supply of consumables, the Mix ability (learned by earning AP as a Chemist) handles revival rather nicely.

(Notice that I didn't even need to bring up White Mage in this post before this parenthetical comment?)

You probably don't need this to be more complicated than Pokemon. In Pokemon, each move has a limited number of PP; the attack costs 2 if the target has Pressure (an ability commonly seen on legendaries) and 1 otherwise.

Also, I don't like durability being consumed by things other than the player's actions. In particular, I prefer the SaGa approach were, even in the games with weapon durability, you don't have durability for anything you don't actively use (so none for armor). Even when using a shield, the cost is once per turn, instead of once per attack blocked (and I note that using a shield counts as your action for that turn, which makes them less useful than in most RPGs).

I remember suggesting to him to change the types to "unsigned" in hope of fixing the issue. i remember him doing something, in an update, but i can't remember what it was. His advice to Zarista, of course, was to knock off the enemy levels so as to stop encouraging players to get to the point that problems arose. Yeah, i remember sending him some info on ths, but programming was new to him at the time, and he was also busy dealing with his web host. I probably still have some of the emails, here. I remember the main reason we were emailing was because of the overflow issue, as well as I believed I had found a potential piracy issue with the addons given how they worked on android. I won't disclose here, though, but his response was interesting.

I think you have an interesting point, there: there's too many games out there where you're the sole character and you can't build a team. Recruiting independently thinking AI needs to be a thing in more games. Part of "RPGs" is that there is a role. A jack of all trades fills a role of 1, but this is indeed an issue with TES games: there's few AI companions, and the ones you get aren't that smart or great. You need mods to improve them. I used to like playing as a God, too, unvtil i started playing games like Vampire the Masquerade: Bloodlines, where I learned that maybe I shouldn't have it all on one playthrough. Challenging myself to deal with the same problems in totally different ways due to class restrictions has contributed to alot of amusmeent, for me. That said, though, those games do a good job of making it so that you can tackle those situations regardless of class, but still in a class specific way.

One problem, however, is that it's possible to escape the durability in TES games, if you just use bare hands, spells (especially in Oblivion where Magicka regenerates on its own), or bound weapons. SaGa Frontier 2 also has this issue, with there being weapons that never break, and spells in party battles not using up durability. (SF2 even has a mechanic that makes it possible to use your strongest spells in every fight; with that said, physical attacks do end up doing more damage than spells, so the usual situation is inverted here.)

As I said before, if weapon durability is in the game, it should either be there as a limiting factor in a way that's strategically significant (a criteria that TES fails to satisfy), or not be there at all.

By the way, one rather interesting use of a durability system: In SaGa 1 and 2, martial arts (like Punch) do more damage when they're low on uses, and the final use does triple damage on top of that. (There's also the Glass Sword, a really powerful weapon that has only one use (except in Final Fantasy Legend, where they goofed when translating the game), and which also shows up in the Ultima series.)

Not even a baby dragon?

SaGa 2 actually has a baby dragon as one of the starting character choices (and that choice has the advantage of allowing you to get a Sprite, which is a good spellcasting monster, early). Furthermore, a female esper (mutant in FFL2) starts with the exact same breath attack that the baby dragon has.

(If you're going to restrict things like breath attacks to certain species, how about making such a species available for the player character?)

I sometimes like putting jack of all trades characters in parties.

Take, for example, the Final Fantasy 3 remake. In this game, Red Mages are actually quite useful throughout, while not making other classes useless:
* They can use some good equipment. This includes Excalibur late in the game. They do have an equipment drought around mid-game, but the ability to use shields can help if you're not using physical attacks (25% damage reduction). No Ragnorak, and stats aren't as good, but still usable as a physical attacker end-game.
* They can cast spells, most notably healing spells. This includes Curaga, which is good enough for late-game emergency healing, particularly if my Devout is busy casting some other spell. Drawback is that they get so few uses of the spell, and don't get the even more powerful Curaja (or Arise, for that matter).

(Certainly not anywhere near as powerful as the Hero in Dragon Quest games, but still a better caster than DQ3's hero before the ultimate spells show up.)

I'll have to address any more posts after this later, as i need sleep.

I'm curious: why would yo uwant to limit flare casting? Wouldn't it be easier to punish them for abuse by making it prohibitively expensive to spam, thus leaving them in the middle of a dungeon without magic? So you have an HP/MP generator, basically. At the very least, you need 1 white mage/summoner/chemist and said blue mage. If the boss bullies the blue mage, you could be in trouble ifyou don't sacrifice another slot for another blue mage. Meanwhile, what happens is the boss bullies your revival character? I, on the other hand, hate those arbitrary limitations. Why is it i run out of thundershocks and can turn around and throw thunderbolts? It doesn't make sense, and just turns into an excuse to go back to the pokemon center. Meanwhile, i think weapon maintainence is a reasonable limitation to a melee character's simplicity. If all you do is buy armor and poke with your sword, it becomes way, way too easy to play vs playing as a mage (which always seems to be the hardest class to play in a game) or an archer. I think the tanky classes have long deserved a bit more complexity and depth. The purpose of damage mechanics isn't necessarily to limit the number of attacks such a class does, so much as offset their economic advantage over mages (who have spells to buy way more expensive than armor, and often mana potions) and archers (who often have to buy arrows). Tanks never really seem to need any maintainence, otherwise, which is why things always seem to be unbalanced in their favor outside of combat.

Well, using SaGa 2 as an example:
* Much of the time, there's no place you can buy them (aside from a shop that becomes unavailable once you turn in the owner). This means you have to be careful about using them.
* When you can buy these tomes late game, they're expensive, and if you use them on small enemy parties (particularly singleton enemy parties), you're going to be spending more money buying replacements than you earn from killing the enemies. Better to save those tomes for bosses, or (in the DS version especially) use them against those chain battles which can have as many as 50(!) enemies in them.
* Remember that SaGa 2 has no MP, so item durability is the *only* limiting factor.
* Also keep in mind inventory space: SaGa 2 has 16 slots of shared inventory (inaccessible during battle), and each character has 8 slots, which need to be shared with any special abilities (like the Robot's OPa/Po) they might have.
* Flare *is* the most powerful spell in the game.

Some options (in Final Fantasy 5):
* Have two revivers.
* Use Phoenix Downs. They're expensive (very much unlike in FF4), but they can be used by almost anyone and will revive at 1/4 HP. (Note that Chemists need these for their full HP/MP revives, however.)
* Cast protective spells on the revivers so they don't get killed too often, and keep healing. (If using a Blue Mage to heal, keep their HP up because of the way White Wind works.)
* Find some set-up that makes the reviver immune to any attack the boss can throw.
* (For situations like Low Level Games or romhacks), give your reviver !Hide (from Bard). Now they can hide, and unless the character decides to show themself, no enemy attack can hurt the character. (This doesn't work so well in battles you can run from, and in the SFC/PSX versions, there's a nasty bug triggered if an enemy attempts a physical attack without a valid target, as well as a softlock if the enemy uses something like Zombie Powder.)

This is actually the issue I have with the (A)D&D magic system, which is seen (in similar but modified form) in Wizardry 1-5 and Final Fantasy 1 and 3 (not 2).