Something about the way +spell damage and healing works has always struck me as wrong.
In the current system, the amount of bonus an individual spell gets is given by a coefficient which is related to the cast time of the spell. Specifically, it's cast time divided by 3.5. This means that all spells get the same benefit from +spell damage over time. You get the same benefit from spamming Flash of Light for 15s as you would for spamming Holy Light.
And that seems wrong to me. Holy Light is the more powerful spell. It should get a bigger boost than Flash of Light. Indeed, the fact that both spells benefit equally from spell damage makes Flash of Light the better spell, in a way. It's cheaper to get the benefit of spell damage with Flash of Light.
Take a look at Blessing of Light. Blessing of Light gives a small bonus to FoL, and a much larger bonus to HL. This feels right. There should be an advantage to using the more powerful spell.
I think that rather than using cast time to determine the coefficient, a different metric should be used. To me, the best measure would be the amount of mana spent per second while casting the spell.
If you spam HL, you are spending 840 mana every 2.5s, or 336 mana/s.
If you spam FoL, you are spending 180 mana every 1.5s, or 120 mana/s.
Eyeballing that says that HL should get about 3 times the benefit in spell damage that FoL gets (on a per second basis). Intuitively, that seems reasonable to me. HL is our power spell, represented by its higher cost, and it seems right that our powerful spells should get more benefit from spell damage.
Blessing of Light tracks this ratio. BoL grants HL roughly 3 times the benefit it gives to FoL.
Because benefit is not linked to cost, +spell damage biases the caster towards cheaper spells. Restoring the link between benefit and cost will make using more powerful spells more attractive.