I would put this in much stronger terms. The theory unambiguously indicates that changing from block reward to txn fees will fail under the current design unless mining becomes monopolized.
We will return to this topic in 30 years, when it will be possible to verify these claims.
For now, stick to the matter at hand. And the matter is, we don't need any fix for a problem that neither does exist yet or it is certain it will ever exist.
My guess is that a single entity will acquire 51% sometime during the next block reward cycle. (that is during the period when the block reward is equal to 25, starting in December of 2012). If the individual is malicious (admittedly unlikely), the developers will likely be taken by surprise and won't have time to save the blockchain. This idea of waiting for bitcoin to die before taking precautionary measures is foolhardy.