Tom Ellis writes:

This is becoming an interesting topic, not - for once - because it's
about pins or tails first, but because it gets into the different ways
of thinking and the assumptions we make. For example, John talks here
about the margin of error for the thickness of a marking gauge pin. But
I never use a marking gauge for this, because a cutting gauge - with a
blade rather than a pin - doesn't have that problem.

A cutting gauge has its bevel towards the fence of the gauge (at least,
all of mine have). When the gauge is set by trapping a board through its
thickness between the fence and the cutter, as distinct from a pin, one
has the whole thickness of the cutter, not half of its thickness as in
the case of a pin, to be taken into consideration where a relatively
high degree of accuracy is desired. Using pin or cutter, it makes little
difference, in this way, one still finishes up with the ends of the
tails and pins proud of the work. As I stated previously, there is
obtained a union of boards with the ends of dovetails and pins
protruding all round by 1/32" (if a pin is used to mark) or the
thickness of a cutting gauge (usually a full 1/16") or something very
near. The "something very near" occurs because few boards are absolutely
consistent throughout in their thickness but the gauge is set, using the
board's thickness at any convenient place, without bothering to take
this into account. Planing the endgrain of tails and pins works O.K.
because of the inherent support given by the joint.

Regards from Brisbane,

John Manners
------------------------------------------------------------------------