I think there are still two definitions of what it means to be self-replicating floating around here.
Until we get them separated and new terms in place to describe them the arguement is basically "IS!", "IS NOT!", "IS" and that isnt' helpful.
How about:
1)self-construction -- when a molecule speeds up or causes the construction of duplicates of itself from anything which is not fully itself.
2)deep self-replication -- when a molecule causes the construction of itself from "simple" components which can be explained by natural mechanisms separated from the molecule in question.
Then we have to argue about what the meaning of "simple" is.