Just a quick comment. I have to agree somewhat with TB here wrt protein replicators (although not really for the reasons he gave ). Basically, one key factor in identifying the first biological replicator is that - for abiogenesis to be true - it must not only be autocatalytic (which the protein is in this case), but also mutable. The underlying problem with protein as first self-replicator is that it can only create exact copies of itself: any addition or deletion of amino acids removes it's autocatalytic properties. It isn't, therefore, mutable. On the other hand, both pyranosyl-RNA and PNA replicators are mutable - beyond a certain point you can add/remove base pairs as much as you want without effecting the self-replicating capability (I think Schleigman went from 4500 bp to 220 bp pRNA over 70 generations or so and still had a replicator). Meaning you can have new features added to the original chain, and hence variation, and ultimately evolution by natural selection. Once you've set up the nucleic acid replicators, coopting amino acids and catalysing their production, glomming on to lipids, etc is just chemistry.
Now actually getting to pRNA or PNA outside a lab is a bit more chancy. Both require pretty stringent conditions. Guess where the creationist designer/god is currently located?
You may be on the right track inre iron sulfide catalysts (or at least templates). There's a group at Glasgow working on an OOL hypothesis based around inorganic templates at submarine volcanic vents (Russell, Hall, Rahman, Turner). Here's a link to their site: The Origin of Life at a submarine alkaline seepage. Earth as a photochemical battery... It's an interesting theory. Maybe you could check it out and see if they're missing something. I think they end up with something like PNA (although they only mention ferredoxins).
With reference to why almost everything has the same chirality, I think you're probably correct. It's just a detail. Regardless of how it started predominating, once homochirality became the norm, anything that didn't match didn't make it. It could even be by simple stochastic processes - sort of like genetic drift randomly moving the frequency of alleles in a small population to one side of the equation or the other. After all, we're talking about the ultimate in founder populations, right?