Not sure whether this should be here or in Knotting Concepts...
The humble bell ringer's knot. A line hanging down from above has a turn twisted into it (a mid-air half hitch) and a bight is taken lower down and is passed up through the half hitch from below, which results in the bight of rope and the half hitch encircling it holding suspended the eye thus created and the working end. The weight of the rope tensions the structure just enough to grip the bight and not enough to cause it to spill, and the bell rope is held up out of the way of whomever is below. From below the half hitch pull downward on either of the legs of the bight and the structure falls apart. This is also true of the leg of the eye descending from the half hitch. The same is generally true for pulling on groups of two of the three descending legs, with one possible exception. But grasp all three descending parts and pull downward evenly and the structure holds.
Having tied the bell ringer's knot, if one passes the standing end (or a bight from the standing end) up through the bight above the half hitch the knot is transformed from what is generally a very insecure structure into the king of knots, the bowline. Passing the standing end up through the bight transforms the simple half-hitch-encircled bight into the nipping loop and collar (which eliminates the spilling to which the bell ringer's knot is prone) of one of the most stable and reliable of knots.
This bell ringer's knot is the basis for the mid-line eye in one version of the trucker's hitch which is often dismissed as insecure. But the nipping turn in this structure, and by extension the structure itself, can be secured without transforming it into a bowline. One way of achieving this is to put a twist in the eye that opposes the direction that half hitch in the tensioned knot tries to spill. In other words, the twist should move the oncoming leg of the eye in the direction opposed to the direction it comes out of the nipping turn. It has also been my experience that a very short eye-loop naturally resists spillage. This method is reliable when executed properly and has the benefit of untying effortlessly when the load is released. This method is shown in this video, although the language is Mandarin. In the video, he also mentions the option of making a round turn, but does not show it:
https://www.youtube.com/watch?v=3zl8EIrHvuQ&feature=youtu.beI have lived in Taiwan for the past 30 years, and it is still very common to see loads on trucks/lorries secured with a modified version of the trucker's hitch described above. The version used here differs from that above in that the nipping turn is not a simple half hitch -- it is formed with a riding round turn. the the rope creating the nipping turn passes around the bight twice. the second turn rides over the first turn, pinching the second turn (the one continuing into the ongoing eye-leg) between the first turn and the standing end of the rope. This, combined with the back twist of the eye creates a very stable pulley/eye/sheave structure for the trucker's hitch. Its behavior when used in a trucker's hitch is markedly different from a single turn or a simple round turn derived nipping structure. I have seen this knot in use securing loads of all sizes. It is being supplanted by straps with ratchet hardware and is becoming gradually less common, but is still a viable option and is chosen in many cases.
This was a lot to go through in getting to my point, which is this: Why not apply one of these/this same mechanism, which adds immensely to the security of an inherently insecure knot/structure -- the trucker's hitch executed with a bell ringer's knot eye structure -- to a the king of knots, which in modern day materials, lacks therein? It seems there is a structural element that does just this -- the riding round turn used in the nipping structure. The "riding turn can be taken in one of two directions, depending on which way the riding turn exits the knot/nipping structure. One has the riding turn (the one on top) leading into/from the standing end and the other has it leading to/from the ongoing eye leg. This mechanism seems equally applicable to right-handed, left-handed (cowboy), and Eskimo (anti-) versions of the basic bowline structure.
Each of double bowline variations that I have played with ("double-bowlines-with-a-riding-turn"? "riding bowlines"? "Riding cowboy bowline" -- I like that) dress into compact, handsome forms and when drawn up seem very tolerant of shaking and cyclic loading, though at this point I have not done anything akin to formal testing. I have not seen any references to the riding round turn device used as part of the structure of a bowline.
JEP