On 8/30/2015 9:54 AM, Peter R wrote:
> Like Daniele pointed out, the greedy algorithm assumed in the paper is
> asymptotically optimal in such a case.

I'm convinced.