I think this was a really useful talk, and well-delivered. I liked how you described the process by which you came up with your current solution, and the iterative improvements you made. Nevertheless, there are some things that could be improved in my humble opinion:
1) I think it would be good to make a distinction between the speedup (due to parallellisation) and other advantages provided by Docker (isolation, testing on different PHP versions, etc.). Running test groups in parallel is probably also possible without Docker (using GNU 'parallel' alone), so it would be good to 'justify' the use of Docker on other grounds.
2) If I read your slides correctly, it seems that the first complete test run (plain PHPUnit, all tests) is actually faster than the last example (Docker, in parallel, with output shown). That feels strange, given the goal of making tests run faster. Maybe the trick you mentioned to speed up Docker container startup time can improve on this.
Thank you for the useful content!
Thank you for the very comprehensive overview of incompatibilities and new features in PHP 7.1. To save some time, you could shorten the treatment of some of the more obscure cases that most developers do not use anyway (octal literals, deprecation of `"$a[-1]"`) in my opinion. Also, it might be useful to show the relevant 'detection methods' for each feature/incompatibility on the bottom/side of the slide where it is discussed.
All in all, this was a really useful talk.
I liked the general content of the talk, highlighting the changes, both breaking and life-improving. Presentation style was also pleasant and relaxed.
It probably would've been better to go into a little less detail on each one, especially with regard to finding the right tool to detect incompatibilies, so we wouldn't have had to skip so many slides.
With that in mind, I would love to have a copy of those slides here to have a peek at the parts we missed.