SunshinePHP 2019
7 - 9 Feb 2019 at Embassy Suites Miami International

I was impressed by the level of real-world problem solving that clearly went into this talk. Amrita, unlike many of us, has experience developing a microservices-based backend consumed on a large scale by clients on many different types of devices. I'm still not completely sold on GraphQL, but she made a great case for it as a solution to the problem of different clients needing to make highly customized queries across objects' fields and relationships. "Convention over configuration" in APIs is still a hard problem, and it was great to hear from someone who found a workable and scalable solution.

I got a lot out of this talk. I can relate to Nara as someone who is in the process of joining the PHP community and wants to get more involved. In particular, her insight that (as an introvert) you get more out of conferences by speaking than by "just" attending really impressed me. Also very useful was the two-axis graph of introvert vs. extrovert and outgoing vs. shy. I personally am more of the classic "shy introvert" but it's interesting to know other permutations exist. Navigating dev team relationships isn't always the easiest for me either, and I liked Nara's concepts to tackle this, such as the "feedback sandwich." One of the better non-technical conference talks I've seen.

This was one of the standout talks: a confident and breezy, yet substantive introduction to serverless cloud computing with PHP. Staying carefully framework-agnostic, Rob used detailed code examples to demystify serverless in a way that made it feel approachable to me for the first time. More than that, he made a great point that PHP is better positioned to take advantage of serverless than other languages, because PHP developers are already used to sharing no state between requests. His treatment of the (very promising) "Serverless Framework" was a little brief; clearly it contains more options than could be covered at this length. Still a great talk. Thanks!

Solid talk, although I wouldn't have gotten as much out of this one without Miro's previous (uncon) talk, "Introduction to Event Sourcing + CQRS." As I mentioned in my review of that one, it drew its strength from detailed code examples and discussion of design patterns. This talk had more scattered case studies and it was harder to follow the narrative sometimes—possibly because Miro was not at liberty to reveal certain details of his previous employers' systems. However, there were still enough nuggets of wisdom throughout the talk that I could piece together Miro's worldview regarding ES. So I'm still giving this one five stars as part of the complete package with the uncon talk (which too few people saw, unfortunately).

Most of all, I appreciate that Miro sees himself as an ambassador for this new CQRS/ES paradigm but without being a hype man for it. He is refreshingly candid about its limitations, and I was able to walk away from this talk with an idea of where ES should be used vs. where it should not. That's more than I can say about any article or Hacker News post I've read on the subject.

I learned a lot from this talk. I've read about CQRS/ES intermittently in places like Hacker News, and I was curious enough about it to brainstorm how we might implement it at work (for some of our data) late last year. But I didn't have a proper introduction into the ES developer's mindset until I saw this talk. It requires different patterns and a different way of thinking to some extent. Miro clearly has a passion for architecture, and his slides took the audience through, not just a finished implementation of ES, but his thought process behind that implementation. It was easy to see what motivated each part of the design.

Very well-organized talk, and with more substance than a simple review of SOLID. (Honestly, though, many of us could use a review of SOLID—this talk made me realize I had violated the Interface Segregation Principle more than a few times in the last codebase I worked on.) I appreciate Brandon acknowledgement that, on the one hand, you should stick to established design patterns, but on the other hand, those design patterns are general enough that it takes some ingenuity to fit them into your particular use case.

Brandon had some harsh words for traits (which I'm generally in favor of), but we had a nice discussion about it afterwards. It's always fun to talk design patterns with someone who's heavily interested/experienced but not too dogmatic.

P.S. Shout-out to the audience member at this talk who said he had an interface at work with 300 methods!

Nicely done...and that's despite connectivity issues resulting in a backup plan rather than a live demo! The digression into why MS is hopping on the Chromium bandwagon also fits super well into the presentation topic, so please keep that in :)

I recall another attendee asking for benchmark performance differences between vanilla JS's querySelector and jQUery's, similar to the comparison you did between a bunch of frameworks on document.getElementById(). Seems like it'd be a decent addition if you can find a like-for-like test, since fewer folks know about querySelector, it has a bit wider use case than getElementById, and honestly it's relatively easier to type (though of course $('.someClass') is easier still).

One other question: I assume you'd let Babel/Gulp pull in polyfills as needed for older browser support, rather than explicitly adding in individual polyfills, hence why you didn't (or maybe I just didn't recall) include mention of the latter in your talk?

All of the PHP in this talk was excellent. But there was only 10-15 minutes worth of PHP and the other 45 minutes was beer brewing. This should have been a lightning talk with the beer removed.

David thank you! This was a great talk on a subject that is often forgotten or overlooked. I'd like to review the content and share it with my dev team. Could you share your slides?

Great information thank you. Could you share your slides for review? Thank you.