What MySQL (and really, Sun) can learn from PostgreSQL

I spent April 17th, 2008 flying to San Jose, Ca only to arrive 30 minutes before my talk, "What MySQL can learn from PostgreSQL", jump in a cab and literally walk into the door of the room I was assigned, "right on time". I think the talk went over fairly well. I opened with the statement, "This is not about MySQL AB, this is about MySQL and the community." I think it help set the tone for the presentation. I didn't want people to feel like I was attacking a profit model or a company. Some items that came up that are not in the talk: * In place upgrades This is a long standing and embarrassing wart for PostgreSQL. No, Slony is not a solution. An individual in the talk asked about in place upgrades for PostgreSQL. They have a database that is 5TB in size. Let's think about that for a second... a database that is 5TB in size. There is currently zero viable solution for upgrading a PostgreSQL database 20% of that size, let alone the full 5TB (if you are a 24x7 shop). Our largest competitors have in place upgrade (MSSQL, MySQL, Oracle). Why is it we don't again? A gentlemen from Sun is currently working on the problem, hopefully he is able to make some progress for 8.4. Another tidbit that I found out while having an excellent conversation on community development with Marten Mickos is that he is currently Josh Berkus's boss. Which means the former boss of MySQL is now the current boss of a PostgreSQL Core member (although up the chain a bit). Lastly, one thing I noticed about MySQLCon is that the conference was focused on specific solutions. The last couple of PostgreSQL events I have attended have been focused on PostgreSQL the server (with a couple of exceptions like ptop). I think its time we reach out to folks like Catalyst, Drupal and Groovy. We need to get people talking about solutions on PostgreSQL, not just PostgreSQL itself. Thanks to HarrisonF and the MySQL folks that wore PostgreSQL shirts so I wasn't alone in the room!