I think Bitcoin will not be a significant long term player in virtual currencies, because it does too much. It provides a pseudo-anonymous virtual currency, a pseudo-anonymous virtual commodity, a distributed transaction processing system, a distributed ledger system, a distributed analog to mining, and the various interesting uses people have come up using the block chain for things other than those things already listed.

For most applications for a virtual currency or virtual commodity, only a subset of those are necessary, and for many applications some of them are undesirable, and for many applications there are additional features that would be good to have.

I think the historical role of Bitcoin will be as a proof of concept demonstrating that a whole bunch of feature, some that were thought to be very hard to do in practice, are feasible to design into a system, but that in practice it will be replaced by several different systems each addressing particular classes of application and optimized for those applications.