I completely agree with this and I am now firmly in the anti-reward-pool-splitting camp. Also, I think that a solution that would work would be if steemit took a more hands-off approach to the development of the steemit.com interface and focused its resources on improving the code. There is a lot of room for improvement in the way that steemd is implemented, and I am pretty sure per cost for development the website is more expensive. I might even say 'hire busy to do it'.