> I really can't see any advantage to this joint.  It's complex but probably
> not much stronger than a single rabbet and may in fact be weaker.

That was my first reaction, but on further consideration, I decide that it
could well be stronger, especially where wood grain is not totally straight
or parallel, and also considering how the normal rabbetted joint typically
fails, if it does.  I thought that a larger glue surface and the grip
further up the joint might add more moment.  My comments are speculative,
though, and entirely based on the assumption of good contact and use of a
strong permanent, waterproof glue that totally fills all voids.

> Because the projection I've marked "X" is cut with the grain running
> across it, it's very weak and doesn't add much, if any, strength to the
> joint.  Any additional glued surface runs around that weak point (|).
> It puts a premium on getting a really tight fit so that the surface (y),
> which can't be clamped, is a sound bond.

Very true, but it does take some stress off the lower piece, just below your
'x'.

> If you do without what amounts
> to a t&g, the corner can be "clamped" with nails or screws, ensuring that
> all the wet surfaces are in close contact.

That is what we prefer, but I would have to see the two corners in
destructive tests to be sure what happens in real life.

> IMO, simple, in this case, is better.

That is my preference, but I still suspect that this corner could offer some
(very slight)advantages and am thus giving it the benefit of the doubt.  I
should add, though that I will likely never make one myself.  (and I used to
manufacture boxes, and have built and been involved in building, using and
selling several finger corner gang dado machines, as well as making quite a
few rabbetted boxes).

Has anyone had any real-life experience over time with the two joints
side-by-side?

allen