Designing a homebrew class, have some concerns

[Sir Edgen]

Hello everyone! After...what, 5? 10? years of ghosting through this website and seeing all the wonderful homebrew classes that people post, I'm finally have enough will to make my own ones, at least start with one. I'm planning to make a class based around characters from one of my favorite GBA game series, Summon knight: Swordcraft story. Also I would like to implement as many features as possible, I don't know if it would be a good thing from Pathfinder/D&D mechanics perspective.
1. In the games characters use different groups of weapons, each of them has its own attack speed. In PF/d&d 3.5 you make the same number of attack per turn with dagger, sword and a giant hammer. How bad would be an idea to assign for each weapon group its own base attack bonus with different progressions? Or should I use something else to embrase this concept, like giving each group different key stat to calculate attack, like Weapon Finesse?
2. How often in your games does weapon durablity come up? Is it a good idea to extensively work with alternative weapon and material characteristics, such as durability and hardness?
3. How much "resource pools" is too much? Aside from being a semi/frontline combatant, my class will have an equivalent of low-caster spell progression and a small set of special abilities of my own making, not unlike martial maneuver system from ToB and PoW. Should I include actual martial maneuver progression or would it be a mishmash?
Thanks in advance. I understand that it would be more convenient for you to just see and discuss these thing after I post the first draft, but I don't want to feel too much shame, if even only in my head.

[Morphic tide]

Point one: The differing BAB progressions aren't a mechanical problem per se, but the d20 system's habit is that the "small, quick weapons" pick up per-hit damage on lower BAB as they're used by the Rogue types, and changing your actual on-sheet BAB messes with prerequisites hard. So it's a touch incongruous with the habit of the system, but it's not an outright problem. Sprinkle in a few Warpriest-style bonus feats, where they count you as full BAB even when you aren't (or 3/4ths if you wish), and the main game function issue goes away.

Point two: Weapon durability in the d20 system is a direct derivative of object HP rather than being a uses setup like in most other systems. It's why Sundering is one of the highest jerk moves in the community, because it becomes startlingly trivial to break the things you're meant to be looting in remarkably short order, and IIRC Pathfinder actually specifically mitigated this with adding Broken as an item condition so things aren't entirely unsalvageable. If I'm understanding your inspiration, do not use a discrete durability mechanic, have turnover imposed by item scaling.

Point three: Going by the Warpriest, it's perfectly fine to have a uses pool, bonus feats, Vancian full-list preparation, and class feature optionality on all one chassis. The Alchemist has three separate resource pools, for all I despise the bombs and mutagens being mandatory and independent from the Vancian extracts. The real issue with including Maneuvers is if they form a true second subsystem, as base classes doing that gets very nasty; If the other mechanics modify Maneuver use, even from additional resource pools, then it's no issue.

Additional comment: Check the Armiger from the third-party Spheres of Might, it seems to be very nearly exactly the kind of thing you want with a class that has its subsystem abilities tied to weapons it makes. It is not, however, an intensive item crafter, nor does it much support off-the-wall shenanigans of the Elemental variety, so your idea still has its place alongside the Armiger as being in-your-face magical.

[Sir Edgen]

Thanks for the answer. I'm glad that my ideas can be implemented easier than I suspected. As for the Armiger class, I'v though about it too. Maybe it isn't ideal for my concept, but there are useful mechanics that can help

[Breccia]

2. How often in your games does weapon durablity come up? Is it a good idea to extensively work with alternative weapon and material characteristics, such as durability and hardness?

In my last campaign, I introduced a couple of specially-constructed weapons that were more effective, but could break in combat. I also introduced a system where PCs and smart enough monsters could directly attack the others' weapons. This included special skills, spells and feats that affected this system.

The players never once used them. Maybe your players will, but you might want to ask them first. I didn't.

Weapon degradation is a realistic thing that real-life soldiers had to deal with, and a perfectly good "price" for additional benefits, but it requires extra effort from your players to make it work. Not everyone is interested. If you want to put it in, make it mandatory, and make sure players can't be instantly wrecked by it. A sword that needs to be sharpened every long rest is not bad. A sword that breaks round four in the first combat in the dungeon, is.

[Sir Edgen]

Thanks. Of course I won't turn this class into "calculate weapon sharpness every 5 minutes because in real life weapons go blunt" minigame engine.

[rel]

keep things simple, keep things contained. A unique special ability is easier to tweak and balance than something tied into existing systems like BAB and weapon durability.

Consider granting some sort of bonus for each weapon type e.g. extra attacks or a initiative boost for lighter weapons to represent faster speed vs
powerful standard action maneuvers for heavier weapons.

To represent weapon switching and crafting, consider granting bonuses for switching weapons or taking special 'craft' actions

In my experience more than two resource pools adds complexity without improving the tactical combat minigame.
That being said, don't be afraid to ditch book keeping. A trivial power without limits or book keeping is often better thematically and more fun for everyone at the table.
And for signiature abilities, a straightforward limit (if limits are even needed) is often just as good as something complicated.