I've no idea whether this idea is technically feasible or not, but if it is, I do think it would solve one of Steem's weak points.
The weak point in question is the Troll Attack. As its name implies, the Troll Attack is an attack on the Steemit ecosystem by a troll who has managed to acquire vast quantities of Steem Power (usually via purchase). Such a troll can vote down good original content, vote up pirated crap that can be found on most anywhere else on the web, and generally create a poisonous atmosphere for all.
Why would anyone do such things? Because that's what trolls do.
Many of us remember the days (only several months ago!) when trolls ran amuck on Steemit. That problem was largely solved by implementation of the reputation score a while back. But, reputation alone can't end a Troll Attack when the trolls has accumulated lots of Steem Power (SP). Here's why:
A low reputation score has no impact other than to hide the posts and comments of those with low rep, thereby likely depriving the poster of any material author awards. While this is effective against trolls who care about posting and the associated author awards (which is the majority of them thankfully), it does nothing to offset the influence of the troll who cares nothing about posting or the associated rewards. Some trolls have managed to accumulate vast quantities of Steem Power and use it to poison the ecosystem while never or rarely posting a thing.
My proposed solution (and again, I'm not even sure it's technically feasible) is to change the way voting weight is calculated. Right now, one's voting weight is entirely a function of one's Steem Power. Hence trolls with high SP and low rep can still terrorize the system by depriving great authors of awards and voting up crap. Who are these trolls? You likely know a few of them already, but some may out themselves explicitly by choosing to flag this post.
It seems to me that the viability of Troll Attacks would be greatly reduced if voting weight were a function of BOTH Steem Power and reputation score. So, perhaps voting weight would be equal to one's SP balance multiplied by one's rep score. Or, if we really wanted to weight the importance of rep, then the the SP balance multiplied by the square of one's rep score. Either would give people with high rep more voting influence per unit of SP than those with low rep. As a consequence, folks with high rep would have more say over the distribution of author awards and could more easily counter the poisonous votes of trolls.
I'd love to hear the community's thoughts on the viability and advisability of such a change.