"40-60fps on phones, tablets, PCs..."
>klicks a button
>site doesn't react for 2 minutes
>laggy background animation
>on i7 2.7GHz Laptop (AC adapter plugged in)
Either I'm missing something, or this is absolutely ridiculous.
The performance problem is because of the abstraction level, we're too far away from the metal with DOM/JavaScript. This can't be solved with JavaScript.
I would try a hard refresh, it's pretty snappy for me (stock Macbook Air and iPhone 5). It's always been really impressively performant, ever since I heard about it some months ago.
Yes, I'm aware of that, but at soon as you interact with the DOM, a canvas, etc, you reintroduce all the overhead that you wanted to circumvent with compilation. And JavaScript without DOM interaction is rarely useful in a browser. For node.js, on the other hand, it sure is awesome. I didn't want to imply that JavaScript itself is necessarily slow, it isn't.
Well, DOM and the canvas elements empirically are somewhat slow, but there's no inherent reason that makes them slow. Besides, if we're talking about something like a game, you can draw the final output to the canvas at a single step which is not the bottleneck.
Would be nice if "About" page worked - just to clear some stuff about framework.
Also, something's not quite right with menu background on Android (Chrome) disappears after a bit. Although, if you click on "menu" button, it's visible, but only until you close it.
I really hope this is successful. But to me, animation performance is really a small problem in the larger problem of deploying HTML as an app. Sure, having something that performs on par with native apps is great. But, both Ionic and WinJS seem to be tackling that problem right now. And, both don't use some esoteric 3D rendering engine and are fully compliant with Angular or anything else you want to use.
The bigger problem, in my mind at least, is gaining full access to the same APIs that are available to native DEVs. Cordova does a pretty good job, but if you go off their core APIs then you are stuck trying to build or figure out plugins, which is less than ideal.
If you visit your own link and then click on the "info" link at the top, you'll be taken to a G+ post explaining that the three.js periodic table demo was an attempt to recreate something the author had seen.
If nothing else it's an indictment of the weird beta-but-not-really-beta that famo.us has spent so much time in, given that this isn't the first time I've seen people assume that it was a three.js demo that inspired the famo.us demo, rather than the other way around.
Yeah, it seems the first post about it on HN was over a year ago. I suppose that's a good enough time for a controlled test release. Will be exciting to see this in the wild.
and https://docs.google.com/a/famo.us/document/d/1DE_26fh9nYbU2x...
Source: https://twitter.com/befamous