A really interesting talk that took me back to my compiler module at University. It did feel like there was a lot of background for only a small number of concrete recommendations for PHP code though.
::golf clap:: amazingly well done.
Great deep dive into compilers and optimization. While a lot of the information was very in-depth, he also included practical tips on how to optimize your code to take advantage of these optimization strategies.
This talk started off promising, but the intro turned into a 15 minute explanation that "your code is slow and you need to fix it" and then didn't explain how to fix the code to make it faster. Instead, the speaker digressed into a half hour long explanation of compilers and an advertisement for the speaker's compiler project. There were a few generic tips on performance improvements tacked on at the end of the talk, but there was no discussion of Quercus, Zephir, Phalcon, nor options for speeding up a PHP site or many of the other items listed in the talk description.
Somewhat interesting but most of it after the initial slides about scaling vs. speed was random details about how compilers work that was presented in an esoteric and impractical way.
That longest part in the middle gave high-level view of how compilers work, then some sporadic details (i.e. here are some HHVM opcodes), but very little that a PHP developer could apply to anything. And people contributing to PHP core or hacking VM's already know this stuff.
Returned very briefly at the end to practical recommendations.
Thus most of the talk wasn't practical for any audience, and was perhaps out of place for a talk with the title it was given.
Presenter spoke clearly and slides were clear.
High quality presentation. Clear, quick, packed with info, articulate, practical info.
I got more out of this talk than any other talk in the conference by far.
Not only does he clearly explain the how of making code better but he also makes it clear why you would make these changes.
I'm not sure if this is good or bad, but his talk was like drinking from a fire hose. One great idea after another.
A few really interesting points were touched on (setting measurable goals, how YouTube metrics are more helpful than Vimeo, etc.) But they were glossed over.
There was also no connection to web design or product design; the focus was entirely engagement in ads.
Enjoyed Mel's talk. Useful information provided
Provided good insight into many micro optimizations that are often passed around like Buzzfeed top ten lists, but there was a lot of reading between the lines to get there. There was also a large disconnect from the talk summary and the actual talk -- not bad, but different.