Although those two post-eye-tiable eyeknots are based on the same bend ( ABoK#1424 / M. A 11 ), they work in a quite different way.
Regarding the "Irish bowline" (2), I think that it can be considered as a simple "locked" "Eskimo" bowline - the second leg of the collar of an ordinary "Eskimo" bowline makes a U turn around the rim of the nipping loop it has just penetrated through, and it is tucked into the collar around the returning eye leg. However, doing this, it goes "over" the first leg ( the direct continuation of the eye leg ) - while
the optimum choice would have been the opposite : our main concern in
any bowline is to squeeze and block the movement of the
second leg of the collar ( the main line of defence against slippage ), and not the first. The second leg of the collar has already a lesser burden to carry, because a significant part of the tensile forces coming from the eye through the first leg has already been "downloaded" at the area of the collar, so, to block the movement of the second leg is always a much easier job than to block the movement of the first leg.
I believe that the second way one can use the ABoK#1424 to form a loop is more interesting, because it generates an eyeknot with a novel appearance. ( See the article by Dick Clements on Knotting Matters, 84, 2004 )(1).
1.
http://www.enm.bris.ac.uk/staff/rrc/knots/IrishBowline.pdf2.
http://igkt.net/sm/index.php?topic=2027.msg14396#msg14396
