Thanks, Dan! That's not bad at all.
Your examples remind me once again that the PG syntax will be changing to the (nicer) SQL3 syntax for inheritance in PG 7.2. It changed briefly in PG 7.1 but got changed back to the old syntax because the PG folks decided their support for inheritance just wasn't good enough to justify masquerading as SQL3 inheritance.
I'm going to go look at the datamodel again in the next few days, I've not done so in about three weeks.