You may think Roux but at first this was a corners first method based on first layer + CLL where the middle layer was solved after the corners, the slowest part of that method, the rest was much like described here (but ELL was always the last step then)...
"Why not solve the middle layer while solving the corners for the first layer? - like, making pairs!!"
Then it became PORT and later PEG (I knew most cases for a while =) but using PEG I found that I was way faster the times I had the pairs in position already before the LL corners was done (named PEG 2, the same as PCMS), so I decided to go for this style and leave PEG (was so many cases so look ahead and remembering the correct alg was slow, CLL I already knew by heart).
// Kenneth 14:46, 31 July 2010 (UTC)