Well realistically, we want to make sure we can make the system as flexible as possible so that spells aren't limited to what Stardock came up with.
Aye, that's how I understood the intention. (and I'd offer to have your babies if it was biologically feasible)
I mostly posted this to point out that the scope of both spell cost/resources and affected target - or multiple targets - should be put on the "flexible" list.
Like a spell/ability that has 2 effects at once with differing scope like:
Targeted city: -10 metal
+
Caster's global resource: +10 metal
An example for a non-standard scope on the spell cost would be
Caster's own HP: -10
+
Caster Faction global Mana: -2
Some kind of necromancer-style HP transfer where it would be useful if the HP would be part of the cost of the spell so the caster
(like an AI channeler) doesn't accidentially kill itself since a spell effect certainly isn't bothered by such details.
So even if AI channelers might use that modded "heal" spell, the matter of the cost wouldn't let them "act stupid" by
killing themselves.
It doesn't automatically require a script for everything "cool" if such things can be worded properly... so the AI can understand it.
Edit:
The only other "issue" I can see is that spells are too straightforward. Too scientific.
One cost, one effect.
For "simple" and small scale spells like healing that is fine and dandy but with big juju, there should be a chance for side effects. Good and bad ones.
Or there could be an entire "Chaos Book", which would always do... more or less what you were intending. =P
My major wish for Spellcasting 2.0 would be that spells can have multiple effects.
<SpellTargetType>FriendlyUnit</SpellTargetType>
<SpellDefEffect>
<EffectName>Protection_from_Fire</EffectName>
<LocalPosition>0,0,0</LocalPosition>
<EffectScale>1.5</EffectScale>
<EffectDelay>0.0</EffectDelay>
<SnapToTerrain>1</SnapToTerrain>
</SpellDefEffect>
<!-- the interesting stuff starts here -->
<GameModifier InternalName="Chaos_Heal_Good_Effect">
<EffectMultipleChoice>1</EffectMultipleChoice>
<EffectTriggerChance>0,8</EffectTriggerChance>
<EffectTarget>Target</EffectTarget>
<ModType>Unit</ModType>
<Attribute>AdjustUnitStat</Attribute>
<StrVal>UnitStat_HealthRegen</StrVal>
<Value>5</Value>
<Duration>3</Duration>
</GameModifier>
<GameModifier InternalName="Chaos_Heal_Bad_Effect">
<EffectMultipleChoice>1</EffectMultipleChoice>
<EffectTriggerChance>0,2</EffectTriggerChance>
<EffectTarget>Target</EffectTarget>
<ModType>Unit</ModType>
<Attribute>AdjustUnitStat</Attribute>
<StrVal>UnitStat_HealthRegen</StrVal>
<Value>-1</Value>
<Duration>3</Duration>
</GameModifier>
<GameModifier InternalName="Chaos_Heal_Side_Effect">
<EffectMultipleChoice>0</EffectMultipleChoice>
<EffectTriggerChance>0,005</EffectTriggerChance>
<EffectTarget>Caster</EffectTarget>
<ModType>Unit</ModType>
<Attribute>AdjustUnitStat</Attribute>
<StrVal>UnitStat_HealthRegen</StrVal>
<Value>-3</Value>
<Duration>3</Duration>
</GameModifier>
This is an example of a "Chaos Heal" spell, having 3 different effects.
Chaos_Heal_Good_Effect and Chaos_Heal_Bad_Effect have the "multiple choice" bit set so one of them is always selected based on the assigned chance.
There's an 80% chance for a good regen spell but a 20% one for... additional bleeding instead of regen.
If I wanted a multiple choice spell that can completely fizzle, I'd make one of the choices a 10% chance effect that does absolutely nothing.
Chaos_Heal_Side_Effect does not have "multiple choice" set so it can always fire, only depending on the assigned chance.
It also has a different target than the "normal" regen spell itself.
There's a 0,5 % chance that the spell works normally but additionally backfires on the caster as well, giving the caster negative regen for a bit.
If the first 2 effects did not have "multiple choice" set then all 3 effects could potentially fire at once.
With a framework like that we could have interesting magic.
There could be a higher level Greater Heal spell that isn't that much more powerful (therefore easier to balance) but does not have the chance to fizzle like the low level version.
"Better" spells would not automatically have to be better in quantity but could be better in quality or have additional effects.
The Greater Heal could add a 2 turn regen or armor buff, for instance, or even invulnerability (massive armor) for one turn.
Magic could be wild and crazy and not like something from a physics textbook. One action, one reaction. Boooooring.
