Does anyone really need MicroPayments?

Many cryptographers have tried to develop special technology for transferring tiny amounts of value; the theory being that the computational and/or administrative costs of other payment schemes render them unsuitable for small value transactions. In this panel we discussed two major questions: firstly are the existing systems really not useful for small values and secondly might other models such as flat rate or subscription systems be more suitable anyway, and be possible without the need for small payments?