I've just managed to get slony1 to build on my system (MacOS-X 10.3.3) 
... first I had to get postgresql ported to build libpq correctly so 
slony1 could be linked with it.

However, handling of schema changes is essential for me, and I can't 
see how to do that in slony1 right now.
Looking at the concept document, the section (1.3) on managing schema 
changes is encouraging but vague.
I can accept the lack of a transparent mechanism to handle schema 
changes as long as the actual mechanism
is easy enough to use... but there is no detail in the document.

Is there support or this already present which I've missed in looking 
at the source code?  If so, where should I look?

If not how is the process to 'execute SQL scripts in a controlled 
fashion' intended to work?
Presumably, the idea is that slony1 will provide a new function to 
which an sql script can be passed as the parameter, and the replication 
engine will run the scrit locally and force data synchronization to 
occur immediately before this script is replicated in order to ensure 
that the script applies to an identical set on the master and slave 
machines.

Would slony1 be intelligent enough to analyse an sql script and perform 
changes to the set configuration accordingly.
eg.
if a table is dropped, it should be removed from the set
if a table is created, it should be added to the set
if a column is altered ... would the table containing it need to be 
removed and re-added?

This would be a lot better than the application having to explicitly 
instruct slony1 to make such changes itsself, but I don't know if such 
analysis of the sql is practical/possible.