Put a lot of emphasis on how to move from anything else to graph, and specifically to neo4j. The whole bit about Docker didn't seem relevant to the presentation at all. Examples throughout didn't feel congruent with each other. Some were about university data, some where shop data... sticking with a single type of data would make the examples clearer.
Would have enjoyed more exploration of different types of database in each category... mysql vs postgres, mongo vs couchdb, etc. Also would like to know more about how to shard the various types of databases, which was only touched on during QA period.
PS: The description of the talk on the joindin page has messed up formatting.
As far as delivery goes, this was one of the best talks I've seen at the conference. Excellent pacing, quality slides, and even a "live" refactoring demo that was spot on.
The content was also quite good. Some of the examples were a little weak, eg using array_reduce to calculate a sum that could be done using array_sum. That said, the overall philosophies that were covered and the design suggestions and general advice were excellent.
This is definitely a contender for my favorite talk of the conference!
Excellent topic and really good delivery. We all need to be reminded about basic security concerns when dealing with web apps. I would have enjoyed a few more code samples that highlight the best practices that were mentioned.
Really good overview of the history of HTTP/1.0 through HTTP/2. Lots of useful technical information as to what RFCs are current and which ones are invalid. Could have been a little bit more in depth as to how we can prepare to HTTP/2 in PHP land and what will be different.
There was certainly a lot of potential but the intent was a bit lost in having overly busy slides. Nervousness of the speaker definitely showed and hurt the appeal a bit. I would love to see this talk again in a couple of months when it has been polished!
Excellent presentation for creating offline workers. Really good recommended stack (Beanstalk, Redis, and Supervisord) for getting up and running quickly and reliably. A slightly deeper dive on how to use separate queues for separate jobs would have been nice, but perhaps not possible in the time allotted.
A solid talk and a good introduction to decoupling. A bit lacking in concrete examples and could have used more depth on how decoupling facilitates testing. Slides were also a bit dry and could have flowed a bit better.
A great way to end the conference. I would have liked to see just a little bit more content in the slides and a little more polish to the slides in general. The flow and dialog was excellent but something was lost in the visual.
Super interesting talk. The examples were good. I'm still not entirely sure how promises are implemented with regard to generators. A little more information to link the two concepts would be helpful. Or maybe I should just go read a promise implementation.
I think some of the container examples should come with a disclaimer not to use containers inside of applications, *except* when dealing with legacy apps. Besides that, the talk was well thought out and should really help people see how to do incremental improvement on legacy apps.
It's probably really hard to do this presentation without a ton of text. I think it would make a really good workshop or tutorial... there was just so much information and code that it was hard to appreciate the end result.