PyConZA 2015: "PyCon Montréal in 30 min" by Petrus Janse van Rensburg

This will be a whirlwind overview of some of the most interesting talks from this year's big PyCon in Montréal. The goal is to boil down the gist of my 10 favourite talks into a single, super-charged talk that gives you a good idea of what went down, but without getting stuck on any particular topic.

Transcript

1. PyCon 2015 Montréal in 30 min Petrus Janse van Rensburg
2. ~ 40 annual Python conferences http://www.pycon.org/
3. • PyCon • EuroPython • DjangoCon • Scipy • PyData
4. • Turorials (2 days) • Main conference (3 days) • 5 tracks • 3100 attendees • Sprints (4 days)
5. http://pyvideo.org/category/65/pycon-us-2015 134 talks
6. 3 x
7. 5 x
9. Keynote Guido van Rossum • Python3, Python3, Python3 • Diversity, diversity, diversity
10. Guido van Rossum Keynote “Yes, you should all be using Python3, and I know you all want to, but I know it’s difficult… So, we’ve decided to give you 5 more years.”
14. Technical Debt - The code monster in everyone's closet Nina Zakharenko “Always code as if the guy who ends up maintaining your code will be a violent psychopath who knows where you live.”
16. Beyond PEP 8 -- Best practices for beautiful intelligible code Raymond Hettinger • PEP8 is great • But, let’s be pragmatic • Focus on pythonic structure not stylistic correctness
18. Test Driven Development (TDD) with Django Harry Percival • Unit tests vs. Integration tests • Writing testable code makes you think differently about the code that youre writing
20. Systems programming as a swiss army knife Julia Evans • Think of your program as a black box • It communicates with the OS via system calls
21. Systems programming as a swiss army knife Julia Evans • strace - lets you see the system calls being made, e.g. “write” or “open” • ngrep - lets you have a look at the raw network traffic on a port • time - for benchmarking • dstat - how much data is being read / written by the OS
24. Python Concurrency From the Ground Up David Beazley • e.g. Twisted, Tornado • One process, that doesn’t get stuck:
 > threadpools
 > co-routines
25. Python Concurrency From the Ground Up David Beazley • Best Practices on concurrency
 > Have simple functions
 > Don’t share state • Avoid concurrency if at all possible, by breaking your jobs into things that can run on mulitple processes, maybe even multiple machines
28. Thanks https://goo.gl/iqPuzZ @petrusjvr