After seeing a few of these kinds of posts over the last few, ahem, years, I am going to take a stab at organizing a Los Angeles based Hacker News group. Let's start really small. Send me your email address if you are interested in
A) being on an email list
B) getting together in person (only one meeting for now)
Hmm, I don't think I understand the diference. Aren't you both in the capistrano case and the 'span codebase' example adding a table for example, so wouldn't both cases require repeating sql sequences?
Although maybe what you are saying is that you can make 'dumb' sql changes which break compatibility with past code vs 'smart' sql changes and code changes which are backwards compatible?
You can't really trust anybody when it comes to science relating to diet. ANY scientific result would have a party that would potentially benefit from it.