Speaker Interview with Robert Haas

Could you briefly introduce yourself?
I am a PostgreSQL Major Contributor and Committer, and I’ve been involved in the PostgreSQL community for about 11 years now. For most of that time, I’ve worked at EnterpriseDB. Before that, I was a user of PostgreSQL for about 10 years.
How do you engage with the PostgreSQL Community?
I read and write a lot of email, and work on a variety of development projects. Some years I’ve focused heavily on patch review, but at other times I’ve worked on very big features, such as parallel query a few years ago and, now, undo.
Have you enjoyed previous pgconf.eu or FOSDEM conferences, either as an attendee or as speaker?
I have not been to FOSDEM, but I have attended PGCONF.EU twice before.
What will your talk be about, exactly? Why this topic?
My talk is about how PostgreSQL represents a Plan internally, and how it is similar to, but sometimes different from, what EXPLAIN prints out. I hope that it will help people to understand how the planner works a bit better, and maybe encourage people to do some work to improve it, if they are developers.
What is the audience for your talk?
People who are interested in learning more about the internals of PostgreSQL and specifically the planner, either because they want to hack on it or just because they want to understand better how it works.
What existing knowledge should the attendee have?
It would be a good idea to have looked at some EXPLAIN plans in the past and have a general understanding of what they mean. If this is your first time looking at EXPLAIN output, the talk will probably be fairly confusing (but you’re still welcome to come, of course!).
What is the one feature in PostgreSQL 12 which you like most?
I’m really excited by Peter Geoghegan’s work to improve btree indexes. It will be really interesting to see what the impact of those changes are once people put the code into production.
Which other talk at this year's conference would you like to see?
Andres Freund is speaking about PostgreSQL’s I/O subsystem and some of the problems it has. I always enjoy his talks, which are typically the result of a lot of really deep research on whatever the subject is. It will also be interesting to hear Heikki talk about Zedstore and Amit Langote’s thoughts on how far partitioning has come.
Which measure, action, feature or activity would—in your eyes—help to accelerate the adoption of PostgreSQL?
I think we need to look harder at how to make PostgreSQL easier to administer. Generally, I think the experience of using PostgreSQL once you’re logged in and have a psql prompt is very good. However, trying to configure the system for monitoring, replication, failover, backups, and so forth can be pretty challenging and many users don’t want to deal with the hassle. Also, when something goes wrong with the query planner or autovacuum or something, figuring out how to solve the problem can be tricky. We need to look harder at how we can make that stuff “just work,” with as little need for human intervention as possible.