> Hi,
>
> What would be the procedure for directly using slony's database functions
> rather than going through slonik? Is this a Bad Idea?

No, this is intended to become a Good Idea.

> I haven't dug through the code yet. If that is the best way to learn the
> ins
> and outs of doing this, where would the best place to start be?

I'd suggest reviewing the schema documentation, which basically embeds all
of the tables and functions in a highly linked HTML document.  It's in the
"howto" directory, which may be a bit of a misplacement.

The functions are quite readable, and you can quite well figure out what
they're for.  They come in pairs, a "public" one that generates an event,
and then an "_int" internal one that does the real work.  Thus, the public
one causes the configuration change to be published across the set of
nodes; slon then tells each node when to invoke the internal ones.

There's not yet any public "objective wisdom" as to how to use them; if
you fiddle with them on test systems, hopefully you can generate some of
that body of wisdom :-).

There's a few notes in a "helpwhatdoido.txt" file in there that point to
using functions for things that slonik doesn't yet support.  If you find
either good things to do or bad things to do, I'd be more than happy to
add more notes.

As I have told a couple of people even today, "more documentation is a
good thing," and I'd encourage you to add even just a bit of wisdom to
what's there if you can.