For our last experiment with structured text, let's have a look at some
subtleties that arise from Harmony's rather simplistic treatment of lists.

First, observe that, if we change the list structure only in one replica,
there is never any difficulty.  

Try adding a line of lowercase characters between "one" and "two" below in
just one of the replicas and see what happens when you sync.

Next, note that if lines get added to both replicas IN THE SAME WAY, there
is also no problem.

Try adding the same new line to both replicas between "one" and "two".  See
what happens.

However, things are not quite so nice if both the same list in both replicas
is edited.  Try adding another new line to both replicas between "one" and
"two".  Moreover, add "FOO" to the line "three" in one of the replicas.  See
what happens.  (Probably less than you expected.)

Delete the "FOO" and re-sync to repair the conflict.  (The "<MARK EQUAL>"
report signals that Harmony has detected and remembered that the replicas
are now equal.)

Finally, try adding another new line between "one" and "two" in one replica
and add a new line at the very end in the other.  Resync.  Were you
surprised?  (If not, try it again!)  

The lesson is that Harmony should be used on list-structured data only with
great care.  When changes are non-overlapping all will be well.  But when
the same parts of a list structure are changed in both replicas,
counter-intuitive behavior can result.  We are investigating how to address
this issue.

This ends the structured text demos.  If you want to play some more, there
are demos to be found in most of the other subdirectories of /examples too.
The address demo is well documented, so that is probably a good place to
start.

* HEAD
one
two
three

