Talk comments

Jos Elstgeest at 13:39 on 7 Jul 2016

Enjoyed the tutorial, really liked the switching between theory and practice

Peter Bowyer at 11:32 on 6 Jul 2016

The best talk about tech and our communities I've seen. Highly recommended - there's a video on YouTube.

Marian Marinov at 12:28 on 2 Jul 2016

Personally think was the best talks in the conference, from the ones I visited that is. But really wanted to know more of how the new infrastructure looked like. It had so many interesting points but we couldn't hear how they internally managed to do that. The time constrain was a problem to do that, I guess.

Thanks everybody for the comments!

@Youri I'll make sure I'll emphasize that in future iterations of this talk - refactoring the old system was given a serious try (roughly 3-4 months) and did not produce (significant enough) results. Based on that, and a few other factors, the decision was made to strangle & rebuild.

@Stephan the problems mostly centered around the docker engine / client, most of those issues have since been solved (I believe we were dealing with docker 0.5 / 0.6 at that point).

Good talk. Well presented.
I liked the different examples about SQL injection. I knew the basic ones, but looking for an item that does not exist and then UNION with some other table to get data of the database that's not supposed to get out was ingenious and rather shockingly simple.

As others have said the title of the talk didn't fit with the content.
Also, what the speaker described is one way of doing things (for example one could also use upstart instead of supervisord) and it would be nice if they were presented as such. You presented a nice recipe, but not all tools may be applicable to all use cases.

The idea was nice, but the walk was rather dry. I think it would be better to have less antipatterns and have some more elaborate examples of the ones you do keep.
Also, Dimitrios was talking way too fast. Speaking style was good, but just too fast.

It was interesting to hear that UUIDs are not completely random, and what the differences between the different versions are.
I also liked the part about how a lot of configuration can be swapped in ramsey/uuid to customise UUIDs for those who are so inclined.
Interesting talk, well presented.

Good talk about an interesting library. I'm not why I would ever need to modify code before it's actually loaded, but when I do I'll make sure to use BetterReflection!

The theory about RabbitMQ, how it works and how to use it, was interesting and explained well, but the practical part of the workshop was rather boring too be honest. It would be nice if there were a real goal we could have implemented, instead of a basic hello world like example.
I realise that the setup of the speaker wasn't working as expected but that should not have kept the attendees from implementing the original thought out task.