This is the second of my PyCon talks on performance. This talk was aimed at showcasing a few interesting problems that I'd encountered over my career, and showing off what those problems looked like in various tools.
If I'm being honest — and why not be honest, right? — I didn't think this talk went as well as the others. It required a lot more experimental setup, and the setup took a ton of effort. That took away from the talk, which could have used some sharpening.
Still, it's not a total failure. As delivered, it was much more modest than it was conceived, but it showcased some common performance problems and how to identify them.
I love this talk. It's probably my favorite of all the talks I've delivered. It talks about how most writing, and especially code, is actually lots of micro-revisions. It then extends what we know about revision to discuss code reviews and technical debt.
I gave this talk at the very first PyCaribbean, which was a fantastic conference. It gathered a bunch of curious minds and interesting characters. I'm not sure if it was the venue or the talk that contributed most, but I had tons of hallway conversations about it and lots of questions over the subsequent days.
I was so excited to deliver my first PyCon talk. I mean, I love Python, and I love PyCon. I'd been allowed to attend in maybe 2013, and the thought of being a contributor was amazing. The talk went pretty well, but I had a ton of support in preparing it and delivering it.
The focus of this talk is on the various sources of information available for identifying and diagnosing performance problems. In many ways, this topic has been treated better and more fully in the subsequent years, but the premise of the talk — that what we now call observability really requires multiple toolsets.
I'm not sure how I ended up writing and delivering this talk, but it was my first real presentation to people outside my business. My jokes landed, I only tripped over my tongue a handful of times; it went well.
The topic is really dated now: using Django & Bootstrap.JS together. Both things got easier, and nobody's picking Bootstrap for new projects now anyway.