The basic premise of capping crit in some way to encourage build diversity and tone down how out of hand crit damage and proccing has gotten is something I can support.
However, the proposed cap of basing maximum crit damage by the weapon damage would definitely cause problems; as Techies pointed out, damage disparities between runed versions of the same skill would definitely cause lower damage runes to become even less effective.
The naive fix to that problem would be to cap crit damage for all runed skills based on the
unruned version of the skill, but you'd still be left with cross-skill balancing issues.
Another possibility could be introducing a hidden stat, such as proc coefficients, that would cap crit damage bonuses for that skill. In theory, this might be easier to balance in the long run. However, I am also on basic principles opposed to hidden stats on skills, and am wary of such approaches to balance.
That would be somewhat redundant though, since crit chance has proc coefficients already.
Lets say you do 10k core damage and have 200% crit damage.
Skill number 1: Does 100% to one target (100coef) = 20k crit damage
Skill number 2: Does 25% to target (25 coef) = 5k crit damage to each target (x4) = 20k crit damage.
So in that aspect crit damage is already balanced. I'm not saying capping crit based on weapon damage would fix the problem immediately, because it won't. It would just give you a better base to balance the game from, you'll still need to re-adjust the numbers.
Lower weapon damage skills would be spammable and have lots of utility, benefit from crit chance etc.
The 145% of whirlwinds base for example might need to be buffed a little to 175%-200%.