jsantell

I use Firefox on Android as my main driver (former Mozilla eng) and I haven't had a crash, but have noticed sometimes videos don't load. On Chrome too sometimes, but usually the issue is with the lazy loading of images (Firefox lazy loading TBD https://bugzil.la/1542784).

I've been avoiding adding any JS, but interaction observers may be necessary since I haven't been happy with how any browsers handle 5 or so muted webm videos.

Was not aware of #lang lindenmayer, very cool.

Some of the other, older language environments I've found are the C-based CPFG, C++-based LPFG, L-Py for python

http://algorithmicbotany.org/lstudio/CPFGman.pdf http://algorithmicbotany.org/lstudio/LPFGman.pdf http://algorithmicbotany.org/papers/lpy.fps2012.pdf

Chromium-WebAR is a similar project, but backed by Tango, and currently exposing a different experimental Web API than WebARonARCore and WebARonARKit. Soon we hope to update Chromium-WebAR (with a rename to WebARonTango) to have API parity with the ARCore/ARKit-backed browsers

https://github.com/google-ar/WebARonARCore

release days are fun

The dominator tree view[0] is one of the first things being worked on after release.

[0] https://bugzilla.mozilla.org/show_bug.cgi?id=1149367

command+opt+i opens tools on OSX

[member on the devtools team] A strategy and implementation for source maps across all tools is something we're actively working on. It's frustrating for us as well when working on side projects, if that is any indication of how much we'd like to get source maps everywhere

in Aurora (Fx39), profile/time tools create profiles and marks on the timeline, but nightly (Fx40) has a combined tool and the profile/profileEnd commands landing soon

watch/unwatch are from Object.prototype, confusingly enough!

time/timeEnd, profile/profileEnd all work in Firefox up to Aurora -- current nightly has a new performance tool that will support profile/profileEnd in a few days, and will most definitely have it before uplift to Aurora https://bugzil.la/1077464

for fun bug tracking in the web audio tools, arewedubstepyet[0].

[0]http://arewedubstepyet.com/

I don't think roads, telephones and mobile networks are an apt comparison -- it's more about the freedom of making your content freely available, rather than the actual physical network.

A better comparison might be the ability to distribute your own newsletter or pamphlets on the street. You can buy a domain name and a hosting provider and have your own personal page to share, whether it be funny cat pictures for your family, a personal blog, a venue for your own artistic endeavours, or political dissent.

Why should it be regulated who gets to put their own mark on "the web"?

Also can install it on Firefox Desktop

that would be.. tricky. almost impossible to generalize, I'd imagine, but who knows

Someone has gotten it working with the soundcloud api to pull in songs from soundcloud and visualize it, getting an example up on the repo :) or do you mean natively on the soundcloud website? (something else I'd love to do!)

In Chrome, the Web Audio API is used -- not using an audio tag. As each chunk of the buffer comes in to be played and decoded, it's processed for spectrum and waveform data: https://github.com/jsantell/dancer.js/blob/master/src/adapte...

Great, will let you know :) although I need to do research on when/if/how the Web Audio API is hitting mobile Webkit browsers..

I'll look into more webkit browsers -- I know on my Nexus S webkit browser, it has awful performance, but the audio plays. If the webkitAudioContext/AudioContext isn't found, it falls back to the Mozilla implementation, which is an extension of the non-fancy HTML5 audio, which means it will play audio, just be unable to read the frequency data from it -- and while the glowy balls move around independent of that data, there'll just be no boom and shaking due to absence of the frequency data :)

Now to build Chromium and try on vanilla Webkit..

Now fixed in the source -- Chrome Canary v21 has 1 audio channel where earlier Chrome has 2: https://github.com/jsantell/dancer.js/issues/8

Noted, thanks!

Both Mozilla and Webkit implemented different audio (Audio Data/Web Audio) APIs -- wonder what BlackBerry Dev Alpha uses if any.. needs a better 'not supported' message

Awesome! Just made a ticket for this: https://github.com/jsantell/dancer.js/issues/8

And of note, while this is using Three.js, it's just using a CanvasRenderer, since FF has problems with loading particles (without a system) in WebGL. Made a ticket for it since this will improve performance as well https://github.com/jsantell/dancer.js/issues/6

Out of the box, no, as this is frequency data rather than time-domain, but could possibly be configurable in Webkit browsers. For an easier solution, maybe something like: https://github.com/katspaugh/wavesurfer.js

Will incorporate volume and start/stop controls. However, a dubstep slider may not be an option

What version/OS of chrome?

The library is leveraging FFT Corban Brook's dsp.js FFT -- while it's not beat detection in the sense that it's able to tell a track is 140 BPM, you can specify a minimum threshold of a frequency's amplitude to register as a beat, and then that decays until the frequency's amplitude surpasses both the minimum and decay rate -- all configurable