Starting the story from the beginning:
(1) RCT starts out with 1 GPSE in its northwest corner, 2 GPSEs in the southwest corner, and 1 GPSE in the southeast corner -- all roughly the same distance from the epicentre (1 parsec) but the precise positioning of the GPSE in the southeast is what determines which RCT bits will be read.
(2) The RCT construction arm builds a Goucher-Grankovskiy construction arm, a.k.a. the DBCA. (See Pavgran's recipe, which is also embedded in calcyman's most recent recipe.) Future signals from the RCT construction arm will now be routed into the DBCA.
(3) Using slow^1 gliders (direct from its construction arm) the DBCA builds a block and a ship (or just three blocks) to stop the incoming GPSEs, and also stop some escaping gliders from the crash.
(4) Using slow^2 gliders, the DBCA builds the far southeast end of a pseudo-BSRD unit (see here.)
(5) Using slow^1 gliders, the DBCA builds the epicentre end of the pseudo-BSRD, including the new p8 reflectors and a seed that produces correctly-timed gliders to shoot down the old p8 reflectors as shown in the previous link.
(6) Using slow^1 gliders, the DBCA builds the seed that removes its elbow block near the epicentre and creates one at the southeast corner -- with the help of a 90-degree one-time turner already built in the southeast as part of step (4).
(7) Using slow^1 gliders, the DBCA builds a seed for an "epicentre elbow" for itself.
(8) Using slow^1 and slow^2 gliders as needed, the DBCA adds a one-time turner to one of its own signal paths, blocking off the fourth option for signal output, as described here. It temporarily stops using that type of glider in its construction recipes.
(9) The DBCA triggers the seeds described in (5) and (6) with a single glider. Future signals from the RCT construction arm will now be routed through the pseudo-BSRD and back into the ECCA eight aeons later -- see (13) below.
(10) The GPSEs crash into the block created in step (3), leaving a big mess but no escaping gliders.
(11) With data returned through the pseudo-BSRD, the DBCA sends signals that will eventually (four aeons later) use 3-of-4-limited slow^1 gliders to clean up the southeast GPSE launch ash blob and the pseudo-BSRD reflectors, and build the SE Corderabsorbers.
(12) The DBCA sends a signal on its unused lane triggering the OTT and producing the local "epicenter elbow" from steps (7) and (8). The DBCA can now use all four of its outputs again.
(13) Using slow^1 gliders, the DBCA builds a second construction arm, the ECCA (Extreme Compression Construction Arm), described here and in more detail here -- a safe distance south of the DBCA, facing in the opposite direction. The ECCA includes self-destruct circuitry for itself, easily triggered by a single glider.
(14) The DBCA builds and triggers a one-time switching circuit to divert future bits from the pseudo-BSRD to the new ECCA.
(15) The ECCA shoots down all the non-repetitive ash created by crashing GPSEs in step (10).
The rest (in purple below) is mostly copied from calcyman's list here, but with some changes related to where the ATBC is constructed and the details of GPSE ash cleanup.
For the third item below, as calcyman mentioned on Discord, the "decent hole" can best be created by firing the two entire Corderfleets heading NW and SE -- even before the NW Corderabsorbers are built. The "reach-around" ability of the DBCA and ECCA mean that the gliders doing the NW GPSE ash blob cleanup and Corderabsorber construction can perfectly well be sent later than the Corderships -- the gliders will pass the Corderships long before the Corderships can travel a full parsec (which takes twelve aeons for Corderships, but only four aeons for gliders).
The same is technically true of the SE Corderabsorbers -- they could be built later than the Corderships they will catch -- but the SE Corderabsorbers will in fact already have been built in step (11) above.
- Using northeast slow gliders, cleanly destroy the DBCA and any other nearby things (no ugly meteor shower needed, because the ECCA is doing the destruction of the DBCA).
- Toggle so that all future slow gliders from this arm are southwest.
- Clear a decent hole in the ash at the epicentre.
- Build the SW bank of Corderabsorbers.
- Use some slow^2 gliders to create an extra elbow at the northwest (for later construction of the NW bank of Corderabsorbers).
- Fire all of the Corderships for cleanup.
- Build a one-glider seed for the ATBC.
- Delete the current elbow so that the rest of the recipe will head towards the far northwest elbow.
- Actually build the NW bank of Corderabsorbers (the approaching NW-directed Corderships have already been built, but they're safely to the southwest of the construction arm), one of which sends a single SE-directed return glider.
To avoid needing that last "SE-directed return glider", send gliders to build the ATBC at the southwest corner, safely south of the Corderabsorbers. The last Corderabsorber in the southwest will produce an output glider that triggers the ATBC seed. By the time the seed is activated, all of the rest of the RCT ash in the Life universe will have been cleaned up, and the Corderships that did the cleanup will have been absorbed.
I think that the Corderabsorbers in the southwest will still have to be built with slow^2 gliders (since I think nobody has yet floated the idea of a third bootstrap stage). And I think that under the above model, the Corderabsorber and SW ATBC seed construction could be completed either by the original DBCA or the ECCA, but it might as well be with the ECCA since the encoding will presumably be a little bit more efficient.
- The ECCA moves its construction arm to its self-destruct trigger location, wherever that is, and with the last codon of gliders from the pseudo-BSRD, fires one self-destruct glider which completely cleans up the ECCA.
There are various possible ways to avoid those last four aeons as that trigger glider travels back to the epicentre. One possible way might be to figure out whether the last two sets of signals from the pseudo-BSRD could be arranged to be close enough together to make the ECCA's circuitry fail -- and then use the failure reaction as the trigger for the self-destruction.
If anyone sees details that need to be filled in or corrected, let me know.