I've been hearing that performance isn't a problem for the last 20 years, yet it seems like I still spend a lot of time waiting for computers and devices to respond. I've worked in mobile and embedded (arguably not a small segment) for a while, and every major product I've released has had a last minute optimization scramble. The growth of these spaces (and now wearable devices) is only making this worse.
On the other hand, a lot of these systems aren't that old. I don't think Android is a Rube Goldberg pile of legacy. But Google acquired a company to get the V8 JS interpreter team for WebKit/Blink, and the Android team just created ART to replace Dalvik. These are smart people who've done a lot of performance measurement and optimization. If the VM didn't matter, it seems like a lot of money and time spent for nothing, right?
Well considering Rust I'm guessing Mr K might have different views now.
> Recently, it was brought up on Proggit that Chris Smith's "What to Know Before Debating Type Systems" was no longer online. This is a really great article, and in an effort to make sure it survives, I've grabbed the archive.org cache and am 'reprinting' it here. If you're into programming languages, read this and level up!