Hullo.
I've been messing around with the idea of a casting system buildt around the Spellcraft skill. I've got the concept somewhat done, but I've begun hitting the issues with such a system, and thus I come to you.
The basic idea is that casters get effects of their spells based on how high they roll spellcraft, a bit like an attack roll really. Ex: If my level 1 wizard rolls a Spellcraft check of 9 (a horrible 1 + 8 for his skill) when trying to cast a defensive effect, he will only recieve the basic mage armor bonus (+4). But if he rolled 15-20, he might get a +5 or 6 bonus instead.
The same will be true for damage spells. A fire effect will have a number of dice based on the check and if it's a single target or multiple target, long or short range effect.
The "spells" themselves will be divided into something like schools, so Fire will have it's own category covering fire damage, resistance and the like. Utility effects will be placed in various appropriate categories and cantrips will have their own category.
The schools will have tables for the various effects they can provide.
Castings per day will be limited by a spellpoint system, a bit like the psionics of 3,5, but with far less points. In exchange for lesser points , they regenerate /hour instead of /day.
I'm sure you already see all kinds of trouble with this, but I shall list the ones I find most glaring:
1. The skill level will be wastly different for intelligence-, charisma- and wisdom-based casters. This can be solved by making a "Spellcasting skill", but that would make things a little more complicated.
2. Casters such as Bards, Magi, Paladins and Rangers will have to be limited in some way to what tiers of power they can access at what levels, since they can get their skill as high as anyone else. This might require seperate tables of effect for these classes, again adding complications.
3. The line between Arcane and Divine casting might become blurred, making some classes obsolete, or at last less nessessary for a group if one of the other classes are there already. This might be solved by making the schools more specific, but that will lead to even more tables.
4. The well-known spells will all be lost, or their effects will need to be converted into tables for the new system. This may result in some confusion and a *censored*-load of tables. The easiest way to handle this is to minimize the number of tables, though this will also minimize the number of spells overall.
5. Power level. This will either nerf casters a little, or boost them a horrible lot. This will need to be tested before any valid estimate can be made.
6. A lot of caster classes, Such as Wizard, Sorcerer and Oracle will need to have their class features changed to match the new system.
----------------------
So as you can see there's a lot of work still in this to make it work, and even more to make it work smoothly.
I've considered drawing up a new spellcasting class that will be the only spellcaster, but the effects of losing all of the other casting classes is a mess I'm not quite willing to deal with, not now, not ever.
This isn't about making casters more balanced, but about making magic more intuitive and usable without the player having to plow through hundreds of spells just to make a list of a few.
At the same time I will prefer to keep it simple. It shouldn't be "Rollmaster, the Pathfinder edition", so keeping the number of tables at a minimum is a must.
Ideas are welcome :)
Edit: Casters will not have automatic access to all the different "schools" relevant to their class, but will be limited to a certain number based on their respective level.
I would suggest looking at this: Legends of Sorcery. The rules are written for using Knowledge (arcana or religion). I usually went the same route as you, by making it Spellcraft, but let casters use their prime casting stat.