Terminology for referring to branches during merges and conflicts

[Stephen J. Turnbull]

Ben Finney writes:
 > Barry Warsaw <barry at canonical.com> writes:
 > 
 > > I have to be honest, no matter how many times I work out
 > > THIS/BASE/OTHER it never sticks in my head.
 > 
 > I have no trouble with "THIS" and "OTHER"; they easily (and correctly)
 > map to the two branches I'm already thinking about during the operation.
 > I always have difficulty with knowing the referent of "BASE" though.

BASE is presumably the most recent revision that is an ancestor of
both branches.[1]  A graphical view of the tree should help with this;
without that, I think it's very hard to visualize, with it, it's
trivial.  YMMV but WFM.

Footnotes: 
[1]  Mercurial docs claim that there's a better way, namely
constructing the DAG per file, and using the most recent undiverged
version.  A graphical view of the project DAG won't accurately
identify BASE in this method, obviously, but a good 3-way merge tool
will identify the changes against BASE that are common to THIS and
OTHER.