View allAll Photos Tagged HTML5,
1920x1200 Desktop Background powered by 5apps.com and marie-schweiz.de
nutzungsrecht für dritte ist nicht vorhanden. bei verwertung für talks muss man mich vorher fragen.
This is a desktop supporting HTML and branding MyModernWeb.com as a company who stands behind HTML5 completely.
scale3d() and .style.zoom tricks to make content fit a wide range of viewport sizes. It actually works pretty well.
See also, Building Flickr's new Hybrid Signed-out Homepage on the Flickr Code Blog.
To commemorate the fact that HTML5 has been renamed HTML and the HTML spec can now be considered a ‘living standard’
"Timeline" view.
A few expensive image decodes and resize operations, but the page is quite smooth overall in terms of scroll performance.
I noticed in Safari that I was getting some pretty severe jank (lag) when firing the machine gun in this web-based game prototype, so I started digging.
It wasn't DOM manipulation, className changes or node removal - turned out it was cloned audio resources being freed from memory after playing, and garbage allocation interrupting the game animation pretty severely in Safari (and Webkit nightly) every so often.
Using standard HTML5 audio, you can't do "multi-shot" - each audio object is single-fire and monophonic. Thus when a machine gun fires, SM2 clones the Audio() instance to make a new object that can play independently, with its own timeline and events.
These screenshots are from Chrome DevTools. Interestingly, Chrome was less-affected by this issue despite having notably-larger GC events as well. I'm not sure how long Safari was taking, but it felt like up to 1 second in Safari 6 in some cases.
In the first case, the GC event is for 9.3 MB and takes 25 msec - blowing the 16 msec / 60fps ideal framerate budget.
In the case where audio is muted and inactive, the first GC event to happen is for 3.6 MB and takes 1.9 msec.
As Mythbusters' Adam Savage would say, "Well there's your problem!"
The solution to this is to simply create a pool of sound objects and rotate through them, if the play() rate is such that the sounds would need to overlap. While this will consume more memory (if not shared/reused by the browser) up front, it should prevent a lot of dynamic memory allocation and resulting expensive garbage collection events.
In fact, most sounds (like machine gun fire) are short enough that multi-shot is not even needed.
Maybe those guys were right about writing JavaScript more like C++ after all, static and all that.
with HTML5 for Flash-less audio!
In the web app, select the gear icon at the top right and select Labs.
Since I took this screenshot they have added Chromecast Fireplace Visualizer.
For about two decades, Adobe Flash was the only means of creating motion graphics for digital signage. However, this has changed in the past few years due to the introduction of HTML5. One advantage of using HTML5 digital signage is that it offers creators a universal platform. It can run ads...
digitalsignagepress.com/blog/html5-digital-signage-solution/
Finally released as an experimental feature, a month or two in development: HTML5 Audio() support for my JavaScript Sound API, with SoundManager V2.96a.20100520. (Progress from the prior dev version.)
This is the first time this API has had the capability to be 100% Flash-free, and I have to say, it's a rather liberating thought. It's shown here working on the iPad; it may be playback or wifi that's a bit sluggish to get going. The Palm Pre also works, which is a good indication of where mobile devices are going with this stuff.
There are some bugs and inconsistencies in support across differing browsers and platforms, but I expect those to be ironed out over time. This is part of the reason why the feature is experimental, not to mention the fact that user testing in the real world is part of where I get some of the most, and best, bugs found and fixed. ;)
Also on the Githubs:
This fascinating 10-second clip (sarcasm) shows an example of SoundManager 2 + HTML5 audio stress test pages running side-by-side on an iPad and iPhone. I'm trying to figure out when and what may cause the dreaded "cannot play movie" error dialog sometimes shown.
The process roughly involves new Audio().src = 'foo.mp3?rnd='+Math.random() being assigned, and then at onfinish(), the process is repeated until the browser (or device) barfs with "cannot play movie" after running out of memory(?) or other resources.
The buttons are provided as a way of starting the sound create-play loop again, should it stop for some reason. On iOS4, sound playback must be user initiated and if new sounds are to be created and/or played, they must be chained via the prior sound's finish event.
The point is to try to determine what kind of limits (RAM, number of objects, amount of requests/loads) are eventually being hit and causing the strange error.
The iPad made it to 135 plays in this case, and the iPhone (a 3GS with iOS 4.1) has finally appeared to have crapped out (albeit the sound stopped, but no error dialog) at 279. Starting playback again via tapping a button gave out at 291, no dialog.
With the iPad, once failed, subsequent reloads of the page immediately fail to play audio and show the modal error dialog. The iPhone seemed to clean up after itself and was responsive after a reload of the page.
This photo is licensed under a Creative Commons license. Please credit Rob Larsen with a link to Drunkenfist.com, if you use this photo anywhere. Thanks.
Murray Maloney hosted a lunch BOF around the theme of W3C's HTML5 work,
tooling and community. These are my notes from the discussion when I asked Ian Hickson what would help make his HTML5-editing life easier; photo by TimBL.
Corrections, clarifications from participants are welcomed via Flickr comments or email to danbri@danbri.org (ideally cc:'ing the public www-archive@w3.org archiving list).
Transcribing and slightly augmenting my notes from TimBL's photo of my lunch scribbles:
[[
Editors, more of them.
(there is a list of desired talents somewhere, from Ian).
Testing, QA infrastructure
(discussion of when this becomes most useful/urgent)
Tooling:
track every email, figure out its category, section, related posts, issues, links, ... feedback, ...
Volunteers to help at checkin point, ... documenting rational, links to wiki and issue tracker(s), when the document goes in. Or even when a change was
*not* made (and why).
(TimBL talked about issue/release tracking in Tabulator)
Mailing list discussion: Ian noted that things are split fairly evenly between the W3C HTML list and the WHATWG list. Ian tracks both without preference. There are slightly different cultures and expectations across each. The core HTML5 people tend to now initiate things on the W3C list.
'tool for +1-ing?' --dbaron
Ian: WHATWG has voting on whatwg / issues
Ian: also I'd like a more flexible license on the doc; people want (a) to be
able to copy from the spec into code (b) allow risk of a fork. The possibility
of this happening keeps people focussed. Re license, DanC has action to follow this up, and expressed some optimism.
]]
"For many out there HTML5 is nothing more than an acronym to use during some geek conversation in order to show off some pretending modern knowledge. This is probably what happened here as well at the very beginning of our recent web challenge: "...come on guys, let's buzz!!!"
Mobile HTML5 Development is extremely challenging and exciting but it can also be frustrating if we don't keep in mind problems and solutions never faced in the Desktop Web. Dealing with standards not complete yet, fragmented builds of the same product, facing both weakness and potentials of these portable devices, this is just a slice of what we are doing on daily basis in Nokia R&D.
This talk is about main technical problems, solutions, and goals achieved while we were building our online Maps application and we hope to inspire and give useful tips to anybody that is dealing with building HTML5 apps for iOS and Android."