One Year of Hacking Workshops

I started running PostgreSQL Hacking Workshops just about one year ago, and I've run one each month, except for May, when we had pgconf.dev. Signups are now open for August, if you're interested in joining us for a discussion of Peter Geoghan's talk on Multidimensional search strategies for composite B-Tree indexes, but I'd also like to take a few minutes to summarize where we are after one year of hacking workshops, both the good and the maybe not quite as good. So here goes.

We have covered 11 talks by 10 different speakers, and all of the speakers save one have attended the sessions. This is one area where I have been very pleasantly surprised -- I did not necessarily expect all of the speakers to be willing to join us, but people have been extremely generous with their time, for which I am grateful. Here's the list of talks we've covered so far:

PostgreSQL Optimizer Methodology (Robert Haas)
Walk-through of Implementing Simple Postgres Patch: From sources to CI (Andrey Borodin)
Streaming I/O and vectored I/O (Thomas Munro)
Memory & Disk I/O Management (Andy Pavlo)
Intro to Postgres Planner (Melanie Plageman)
NUMA vs PostgreSQL (Andres Freund)
The Wire Protocol (Heikki Linnakangas)
A Deep Dive into Postgres Statistics (Louise Grandjonc Lienweber)
How Autovacuum Goes Wrong: And Can We Please Make It Stop Doing That? (Robert Haas)
PostgreSQL meets ART - Using Adaptive Radix Tree to speed up vacuuming (Masahiko Sawada)
Fast-path locking improvements in PG18 (Tomas Vondra)

The total number of people who sign up each month is usually between 30 and 40, ranging from newcomers to committers. Most people who sign up also attend, although we almost always have a few who do not. Most attendees are from Europe or the US, although we do have some from China and India. We have had very few, if any, attendees from Japan, even though it is a very important country for PostgreSQL. Many of the people who attend do so regularly or semi-regularly, and this is true both for people who active participants on pgsql-hackers and for people who do not regularly participate there. While it is great to see people come back from month to month, we are not at this point attracting very many altogether new people. That seems like a possible area for improvement, though I don't know how to get there.

I am generally happy with the quality of the discussion that we have on the calls. I don't know whether other people agree, although the amount of repeat attendance probably means that people at least aren't hating it. I worry that people who don't speak up are intimidated by the experienced hackers on each call, and I always make sure to emphasize that we are more than happy to have questions from people of any level, and encourage everyone who wishes to take a turn. I think that some people simply enjoy an opportunity to hear other people discussing these topics, without feeling the urge to participate in the discussion. That's OK, but it can be hard to tell the difference between that and someone who does not feel empowered to speak up.

What is hardest to judge is how much impact the calls actually make. I don't know whether anyone is doing more PostgreSQL hacking as a result of the calls than they would have done without them. Several people have thanked me for organizing them, sometimes more than once, so I do know that at least some people appreciate them, which is perhaps a good enough reason to keep doing it, but of course it would be lovely to have more information on whether and how the sessions benefit those who come.

If there are other things you would like to know about this effort, or if you have feedback on it, please leave me a comment, or contact me via email or Discord. Thanks.