Steps:

(1) Cubeshape: This is standard so no need for explanation

(2) First Block: This is the same idea as in Lin so again not much explanation needed.

(3) Psuedo-block: Here we solve the two remaining D-corners with any D-face edge in-between (Second block from Lin but ignoring the permutation of the edge).

(4) CMDLL: We then use Lin algs to insert the remaining D-face edges and solve corners. The extended Lin algs for 2 D-face edges in U can be found here

(5) Permute Edges: Now we're left with a subset of Vandenburgh's EP step where we force U(a), U(b), Opposite, and Adjacent cases on the D-face. This lowers the EP alg count from 100 to 50. Algs here

Example:

Scramble: (0,-1)/ (-3,0)/ (1,-2)/ (0,-3)/ (-3,0)/ (3,0)/ (-3,0)/ (-1,0)/ (0,-3)/ (0,-3)/ (-3,-2)/ (-2,0)/ (-1,0)/ (2,0)/ (0,-4)

Solution:

Cubeshape: (0,-2)/(-2,6)/(-1,-2)/(2,-3)/(-2,-1)/(-3,0)/

First Block: (3,-2)/

Psued-block: (2,2)/(-5,1)/(-4,-1)/

CMDLL: /(1,-2)/(0,3)/(0,-3)/(3,0)/(-4,2)/

Permute Edges: (3.-2)/ (-3,-3) / (0,1) / (0,-2) / (0,-4) / (-4,0) / (-4,0) / (-2,0) / (5,0) / (-3,-3) / (-3,6)

Movecount: 25 STM (with a cancellation). I did a Lin solve as well which came out to 28 STM.

Notes:

(1) Because we solve left block and force U(a)/(b), Opp, and Adj cases, the D-face case recognition is very easy (it should be able to be recognized regardless of the ADF). This makes EP recognition faster and closer to EPLL's recognition.

(2) This should be more efficient that vanilla Lin since we're saving moves on 2nd block and permuting more edges at once during the final step.

(3) One downside is that , unlike Lin, you can't incorporate Yau-1 or PLL+1 algs. This would mean that this method would end up being less efficient than full Vandenburgh and Lin with PLL+1. I'll do a bunch of example solves and try to estimate where the average movecount falls relative to each method. Thread for that here

It's been a while since I've been on this site, so if this belongs in a different thread/has already been purposed, I apologize.

~Cheers