Im working on a site where i will explain my cage method and fewest moves stuff, amongst other things. I won't give out any link yet as the site so far is only a skeleton: empty pages and navigation.
When it's close to "publication" i will get back with the link to my site.
I gave Pembo link to mcFarren's site not because it's so good, but it's the only site that describes a method fairly similar to my method. if anyone goes through his pages bare in mind that the steps are just like mine, but his algs are quite inefficient, no block usage.
His last layer edges is also very inefficient. As i understood it he does it in 2 stages: "orientation" and "permutation". Of course there is no real orientation of outer edges, only for middle edges on odd sized cubes. My way of doing LL edges is very similar to freecycling, 3-cycles. Track the edges and solve 2 at a time wherever possible.
For the final centers this is almost entirely done with commutators (3-cycles). There are some useful exceptions like (r2u2)*2 for opposite faces.
How i fix parity is too long-winded to explain fully here, but basically i do setup moves to put the 2 edges to be swapped into ULb and DLb positions.
Then i do b (4-cycle in b-layer) and R d' R U2 R' d R U2 R2 (3-cycle in b layer). So 2-cycle(swap) = 4-cycle + 3-cycle. Undo setup moves is intuitive. Centers get distorted, but hey cage=centers last. No big deal!
Btw the first layer is purely intuitive, solve cubies in any order (in the order u see the cubies). Get this step fast before pursuing the later steps ...