PNWPHP 2016
15 - 17 Sep 2016
at Cornish Playhouse @ Seattle Center
Talk comments
I've seen an HTTP/2 talk or two that focused a little more on the spec of the thing, but this one was very applied, with plenty of examples...including code samples...and a ton of energy. Fitting, given that Davey was one of the folks pushing for H2 CURLOPT support within PHP.
Included in the examples were pieces of info thatsimplified HTTP/2 implementation enough for me that I think I might actually go out and do it...particularly now that Amazon has a load balancer product (Application Load Balancer) that's up for the job!
...and now for something completely different!
I attended this talk back at Lone Star PHP 2016, had fun with it then, and this time around it was a bit more polished, and had the same clear calls to action for self-improvement...and that's not just the priming words at the beginning of the presentation talking :)
This talk ended up being a lot about history and the "why" behind Laravel rather than the "what"...and that's fine! Laravel and its stable of products are documented well enough to understand, and there are plenty of resources out there, but as a bit of a laravel skeptic the heavy focus on the why's of decisions made, whether Homestead, Valet, the framework in general (including not-actually-SemVer), or other ancillary products was quite informative.
None of the above makes me a Laravel acolyte, but the added context was definitely appreciated.
As one nitpick, the response to the Great Middleware Signature Debate question felt a bit edgy, particularly given the poop-flinging that's happened in the FIG over that signature.
Solid talk, backed by stats from one of the kings of mobile. I knew about a fair number of concepts in the talk (e.g. testing for low connectivity, AMP, interstitials being de-ranked and mobile optimized/first sites being buffed in mobile search rankings), but there were a huge number of nuggets of information about those topics and others mixed in, on top of a solid Q&A section, that meant that I got a ton of information out of the talk.
Heck, that talk inspired me to throw an AMP plugin onto my personal blog shortly thereafter. Pretty strong action item :)
Pretty solid REST talk. For the sake of time a few pieces were hand-waved over (e.g. OAuth/JWT), pacing was a bit off at times (or maybe it was me becoming distracted about material I already know about :) ), and I'd call HAL/Siren/JSON-API and vnd.error/http-problem close enough to a spec for hypermedia for horseshoes and hand grenades (forgot to mention this as feedback on this talk at Sunshine this year...oops), but on the whole a worthwhile presentation.
Practical, memorable, entertaining, knowledgeable, with a call to action of "you've gotta start somewhere and here' a logical progression to do so." I may have taken a different devops path on my projects, but I appreciate Chris's take and think that it's a solid way to iterate toward an application infrastructure that doesn't wake you in the middle of the night.
Content of the presentation was fine, judged by a person who has implemented and helped orchestrate several microservices in Slim. However a few points needed a little more clarity/correction (JWT resource servers only need the pubic key of whatever generated the JWT, withJson() is in Slim\Http but not the PSR-7 ResponseInterface) and the presentation felt a foot deep and a mile wide (I've been guilty of this myself...see some of my own talk feedback on here :( ). Talk may have been better as a focused deep dive into what makes Slim in particular tick.
Also, lose the .0 in the title; Slim's pretty much SemVer-compliant and we're on 3.4.x now :)
Panel was solid, particularly given that it was put together impromptu. It's great to have that many bright people in the room, particularly when they have strong, well-formed opinions :)
Rapid-fire talk by a passionate, extremely well-informed speaker on optimizing for not-assuming-the-laws-of-server-physics-don't-exist, which happens to be one of my favorite topics. Explained several cases of what's slow/memory-inefficient, why it's slow/memory-inefficient, and what you can do about it, plus a call to action at the end of making your software maintainable/less of a headache for the next person to interact with it. Great all around.
Great topic, solid delivery. No ivory tower pretense. Just examples of how you'd use graph databases, with Neo4J as the example (I mean, he works at a hosted Neo4J provider, so that makes sense) to solve problems that are a real pain in an RDBMS. I think his talk was accessible enough that there'll be plenty of folks using Neo4J for the first time as a result of that talk, and that's great.