[Dan Lehman]

Originally Posted by Ian McColgin

Bad plan. The big thing that causes knoted lines to fail is turning radius and the two eyes made like a squre knot manage that.

It's not so simple as that--turning radius around metal, e.g., turns out to usually NOT
be the point of failure, I think. In some testing of a Strangle Noose (Strangle/Dbl.Overhand
tied around standing part of noose hitch), the break came at the point of the sharp,
1-dia. turn, but in the mainline around which this turn was made! Ditto for Buntline h.?!

In addition, it's hugely more trouble to make up than a nice Carrick Bend with siezed ends and that bend has been well tested as weakening the line by but 15 % to 20%.

Except for the pulling through of the entire length of ONE of the eye-spliced lines,
which might be seen as simple and not so much trouble, the seizing of the lattice form
of the Carrick bend seems more trouble. And, as the OP asks, where is ANY testing
information about this structure? --it's been referred to (the structure, i.e.) in many books,
but I've NEVER seen any hint of actual test data for it (and that for the capsized form
is well below 80% of tensile, though that might be pretty laid-rope-particular data).
And, to my mind, a LOT of the integrity of such a structure depends upon the seizing
--which must serve to effectively make an eye in the lines, splitting the force to the
end part; and if one seizing fails, the structure will spill.

If you go through all the work of a couple of eye splices, why not make a nice short or long splice with will not weaken the lines for the same work?

For having size-adjustble rode, er, modules--vs one hugely long line (less easily
turned end-for-end to distribute wear, e.g.).

I don't understand the OP's intended distinction between some TWO methods:
these are really the same thing, though the Larkshead form I think is inferior.
In any case, the assertion of needing access to BOTH ENDS in nonsense: one need
only one end: place associated eye around the end-INaccessible eye, then reeve
accessed ONE end through said other eye and haul away--QED.
One could try to improve breaking strength by repeating this algorithm such that instead
of a SquaREef-knot-like structure one had the additional wraps of a Surgeon's knot.
But it seems that of concern then is the amount of slip-tightening and related frictional heat!?
--YMMV.

Another method of joining eyes is to employ a short THIRD bit of cordage, with which
one binds the eyes together in any of several geometries. E.g., bring one eye down
through 2nd eye; pull 1st eye through a little ways, sufficient to ... ; bring the sides
of 1st eye out around and over 2nd eye's; then bind these 1st-eye sides with a cord,
making several wraps, and tucking or tying cord ends (little tension will be on ends).
This describes a sort of extended Reef-like structure, which should be quite strong,
easily implemented, and easily untied.

--dl*
====