I really enjoyed PostgreSQL Conference WEST this year. It’s a shame I didn’t have a little more time to explore Seattle, however; it looks like a nice city.
There were a lot of great talks. Scott Bailey’s temporal data talk was excellent, of course — but I’m not an impartial observer, I expressed enthusiasm before the conference began. Another talk that I found very interesting was Bill Karwin’s talk on practical object oriented models in SQL. The talk was well-researched and the points were made very clearly, although I would have liked to see a little more depth. There was some audience discussion, and I’d like to see the points hashed out in a little more detail.
I’m always excited to present the projects that I’m working on, and this conference was no exception. I thoroughly enjoyed discussing Operator Exclusion Constraints, the new feature that I’m writing for PostgreSQL 8.5 that enables “temporal keys”. Slides are here — I’ll write in more detail in another article. I was pleased with the audience participation and questions during the talk, and I hope that everyone enjoyed listening to me as much as I enjoyed speaking.
Slides for my talk on extensibility are available here.
Thanks for your compliment about my talk! You mentioned you would have liked more depth. I’m interested in improving this talk and future talks. What specific kinds of depth would you have liked to see?
I found the discussion about testing very interesting, and your points made a lot of sense. However, people often find other creative ways to make testing easier or avoid code re-use, and I’d like to see those ideas hashed out a little more. For instance, you probably remember some audience questions around that topic, but with only an hour, there wasn’t enough time to really think about all of them.
I believe you did answer the questions, but I didn’t have enough time to think about the question, the answer, and how it related to your material. I’ll wait for the video.
Yes, I can understand your point. The talks were really only about 45 minutes long, and my talk had four main topics: EAV, Polymorphic Associations, Hierarchical Data Sets, and Active Record. So each of those got about 12 minutes. The point about testing Active Record-based models is important, but could only get a fraction of those 12 minutes.
I’m currently writing a book, “SQL Antipatterns” based on the content I develop for conference talks and answering questions online. I’ll make it a priority to go into more details about testing models in the book, and include code examples.