14 Sep: W3C is holding its 1st web developer conference in Seattle! Nov 15-16
If you are a developer or designer wanting to hear the latest news on HTML5 and the open web platform, and your place in it, save the date.
W3Conf: Practical Standards for Web Professionals
2011: HTML5 and the Open Web Platform
We have selected speakers on an invitation-only basis to deliver the most educational and entertaining experience for our audience on the topics we felt were the most pressing for Web developers and designers today.
http://www.w3.org/conf/#presentations [downloads and video recordings]
Below is a rearranged for reading and highlighted copy of the live blog of the conference by Manu Sporny:
W3Conf LiveBlog – Day One [Manu Sporny, Digital Bazaar, Nov 15, 2011]
Ian Jacobs (W3C): Welcome: Contributing to Open Standards
Focus on why developers are key to W3C and the future of the Web. W3C doing standards since mid 1990s – learned a great deal in that time: You have to have the right people involved. Who the right people are changes over time. Consensus building takes a lot of time, but you can’t let it go on forever – there is a balance. Royalty-free patent policy is important. Use cases and real-world developer stories are needed. Tests are needed for proper implementation. Developer buy-in is very important.
Continuing improvements at W3C is important – HTML5 logo, community groups, Unicorn testing framework, training material – useful documentation for developers/designers.
Community groups – important for bringing new work to W3C. This conference is a call for action – for web developers to get more involved in W3C.
Video of Tim Berners Lee (W3C):
Asks people here to have fun. Keep the Web open – use open standards. Do really exciting things – HTML5 and Web Apps – the platform allows you to do things that were only possible via native apps before.
Try to focus on Web Apps – not native apps. Being part of the Web is more powerful than creating siloed apps. Bit by bit – everything that you can do on a native app, you will be able to do via a Web Apps. By making Web Apps – you are helping to keep the Web open.
Keeping the Web open has always been a battle – we have to fight for openness.
Philippe Le Hégaret (W3C): Testing to Perfection
Asks audience to raise their hand if somebody has ever logged a browser bug (very few people raise their hands). “How many of you expect the web to work?”. Laughs from the audience.
Shows a testing example on CSS – asks audience what’s going to happen. Nobody seems to know. Different browsers show a different result… 25% of browsers will not work. This is what we would like to improve. Replaces test DIV elements with SPAN elements… invalid document, but will still render, but not exactly the same.
Is testing to perfection possible? No. Testing is an approximation.
Lots of it can be automated, but a lot of it cannot. We need help testing plenty of combinations – HTML, HTML+SVG, DOm manipulation, CSS combinations. We’ll never get 100% coverage – but we still need to set a goal. Who are we testing for?
We are testing for the people that use this technology – manufacturers, spec editors, content providers, people using the Web… etc.
Manu Sporny (Digital Bazaar): Community Groups: A case study with Web Payments
Doug Schepers (W3C): Shortcuts: Developer Documentation
We’re back with Doug Schepers, SVG, Web Apps, Audio, that will be talking about Web Education. “Who here has used the Web?”… “How many have programmed for the Web?”.
“How many people learned the Web via ‘view source’”? Just about everyone raised their hands. “How many of you still do that?” Most dropped their hands.
The Document Web was pretty easy to learn. In 2000, we started to get the application Web. Media Web, Social Web… you have to learn a lot in order to be able to do these things. People can’t read standards – they’re really difficult to read.
WebMoney came early on. W3Schools, W3Fools.com is a good site. Opera Web Standards Cirriculum, SitePoint, MDN, Google’s “Ground Up” Videos. Tons of other sites – not just HOW-TOs, but help more.
At universities you learn serious programming languages like “lisp”. Laughs from the audience. Web Standards Project is great, InterAct Cirriculum is great – teachers can take these cirriculumns and plug them into their classrooms.
W3C has tutorials and primers, podcasts and videos. The W3C wiki has a bunch of resources, HTML5 reference guide, CSS reference guide. The problem with W3C’s documentation is that it’s scattered everywhere. Internationalization documents on W3C site.
Web Education Community Groups – launched to focus on learning material, curriculum, outreach, training, international resources. Trying to improve teaching resources at W3C. Why should people get involved in Web Education? Many eyes, pay it forward, reap what you sow.
How to help: write articles, review articles, help write curricula, help translate.
We started the Web as a hobby. We need to change it from a hobby to a craft. Make it easy to learn the basics, need smooth transition from casual development to career. Web developers learn throughout their lives – we hope these resources will persist and keep getting better.
Divya Manian (Opera): 5 CSS Magic Potions for your Layout Troubles from the Future
Divya is a Web Opener for Opera – contributes to HTML5 boilerplate, member of the CSS WG. Arsonist of the Semantic Web. Current state of CSS layout – what to look forward to in the future.
Many people use ‘floats’ for layout – absolute positioning. Problems with floats – they’re not content agnostic – floats require clearing. So, what can we look forward to in the future?
Paged media layout – “@media paged” – allows you to tread HTML as paged content.
Multiple columns – “css3-multicol – column-width: 12em;” – allows you to setup column sizes. Column spans allow you to say that a particular item can span multiple columns.
Regions allow you to flow text content from one region to another, even if they are separated by a large distance.
Exclusions – allow text to flow around the outline of irregularly shaped objects. For example – a large rock with text around it. Lots of options on how to wrap text via CSS.
Grids – standardize a way to do grids inside CSS. Grid templates look simple to start – very complex.
Flexbox – very much in flux – allows you to distribute layouts in a more flexible box layout.
Make sure to use feature detection via modernizr. Subbornella’s tutorials are great. Isotope jquery plugin is great – isotope.metafizzy.io – if you have complicated layouts, use Isotope.
More discussion on the www-style mailing list at W3C. Lots of traffic on the mailing list. Help by submitting bug reports. Divya can be found on Twitter at @divya
Vincent Hardy (Adobe): Web Graphics – a large creative palette
Vincent talking about passions – Web and Graphics. Formal education in distributed computing – worked at Sun on 2D APIs. Worked on SVG. Worked at Oracle on Data Visualizations.
Towards a more graphical, fluid Web. Shows graphical clock via Raphael. Amazing WebGL demo showing shader-use and cool artistic effects.
You have lots of tools – canvas, WebGL, CSS Animations, imperative vs. declarative programming for graphics. Very nice visual effects in canvas/SVG.
Canvas – imperative model – big array of pixels. SVG – declarative model – looks more like the HTML/DOM.
SVG – graphical elements w/ attributes that control how content gets rendered.
Canvas – pixel-based, you write scripts to draw what you want. You create a context, set it to 2D or 3D, then you specify drawing commands (draw an arc, fill with color, etc)
Canvas – more difficult to detect where events happen in your object – you have to write that code. SVG – events are tied to graphical objects. Easier to do event inputs via SVG – but both allow you to have very fine grained control over your graphics with either.
Scripted animation is still in the works – but you have scripting APIs today. Canvas uses scripts for animations. With SVG you can use CSS animation, scripts, and declarative animation with SMIL. All browsers except IE support SMIL. Declarative animation allows you to morph geometry between objects.
Timing – you can time animation events and chain them together via SMIL pretty easily. You can style SVG using CSS, just like any other HTML document.
Multimedia integration –
You can also do video in SVG – Video element embedded in SVG document, you can apply filter effects to SVG – filter effects apply in real-time. Foreground decorations with SVG – looks really nice.
Why two different models? Canvas is low-level, pixel access. SVG is very high-level, nice API, DOM events, etc… but no pixel access, can be limiting. There are good reasons to have both. Graffiti Markup Language – uses SVG and Canvas together.
Frameworks that can help: Raphael.js, Paper.js, Easle.js, D3.js, Canvg.
Canvas support is very good – supported in all recent browsers. SVG is supported in all major browsers, except for SMIL animation and font stuff in IE.
WebGL support is coming, but not quite there yet. Issues with in-line SVG.
What’s coming: Better 2d/3d integration, improved integration w/ canvas/SVG, additional features for SVG 2.0, filter effects and CSS shaders, video integration, etc. Presentation was created with HTML5.
Arvind Jain (Google): Web Perfomance: Making the Web Faster
Arvind is the Chair of the W3C Performance Working group – try to make web page performance better. Web Timing API – “If you can’t measure it, you can’t improve it.” — Lord Kelvin.
Web Timing API allows web pages to get detailed information about how long it takes pages to load. Web Timing API has four areas – navigation, resource, user and performance.
Navigation Timing API allows developers to create a complete picture of how long it takes for a page to load in very accurate detail. Supported in IE9, Chrome, Firefox and Android.
Resource Timing API – timing info related to individual resources on a page like images, scripts, objects, etc.
UserTiming API allows high precision (1 ms accuracy) to measure the execution time of their code.
Performance Timeline API – single interface for accessing all timing attributes.
Resource, user and performance implementations are not entirely implemented yet, but coming soon, around March 2012.
Security and Privacy implications – trying to ensure that people can’t be easily tracked with these new features.
Web Timing stuff is in use – Google Analytics Speed Report, WebPageTest.org, Dynatrace UEM, Boomerang.
Page Visibility API – helps you figure out of the page is visible or not – helpful to know if you need to keep running animations/etc. Ready in March 2012. In use on Google Analytics (count impressions correctly – page isn’t counted if page is opened in a tab that is never looked at), YUI Idle Timer, Google.com Instant Pages loads pages if it can predict if you’re going to click on a link (loads in the background).
Question on Usability Testing and improving the browsing experience. Yes, the Timing APIs can be used to figure out how long it takes for someone to click on a piece of a page.
Question on using image bundles or not. Google performance testing has shown that it depends on the types of images – large image bundling is bad. Tiny image bundling is good.
John Allsopp (Web Directions): Shortcuts: Getting off (line) with the HTML5 appcache [his slides on the web]
Great criticisms of the Web – it only works when you’re online. Not necessarily true.
For Appcache to work – create an appcache manifest file and link it to your HTML file. You have a file with a .appcache extension – appcache file has sections for cache, fallback and network.
Cache section – which resources must be cached.
Network section – which resources must NOT be cached.
Fallback section – which resources should be used for anything that is not cached and when you’re offline.
Gotchas: Appcache manifest must be served with text/cache-manifest. New-ish technology. Cache failure if one resource is not available – must verify your cache. Appcache is cached forever – when developing, don’t use appcache. Only things that are explicitly cached are available – must be explicitly included. There is lazy-caching, other cache controls in use in the browser may cache other files.
Do not develop with app-cache turned on.
If you change the contents of the appcache, it will be refreshed. Use #version 1.0. Some browsers limit to 5MB-10MB.
Appcache is supported in all modern browsers.
ManifestR creates a manifest file for any web page you visit – helpful for starting with AppCache development.
Rajesh Lal (Nokia): The N-Screens Problem: Building Apps in a World of TV and Mobiles [YouTube video]
7 Key challenges for the N-Screens problem.
Four screens – mobile smartphone, PC, tablet, Internet-connected TV.
Design Problem – N Screens – Mobiles, screen size 2-4″, viewed from a distance of one foot. Not reliable, finger input, highly interruptible. Very dynamic environment. Tablet – 7-14″ screen size, input touch, not as interruptible. PC – screen size 12-27″, very focused use. TV – screen size 25-65″ screen size, 10′ viewing area, total immersion, D-pad remote.
Native vs. Web App
Native, pros: each screen has an SDK, great for a specific device, robust, device APIs. cons: steep learning curve, not scalable. Winner: WebApp solution.
History: Old HTML, Server-side web apps and CGI, Rich Interactive Apps – Flash, Silverlight, HTML5 and now WebApps.
Seven key challenges
- How do you detect features? Device detection (bad) vs. Feature detection (good). Modernizr and jquery.support help with feature detection.
- How do you detect screen layout/orientation? Media queriesare great – you can query based on resolution, based on device aspect ratio and based on orientation.
- Graceful degredation? moz-* ms-* o-* etc… you can use CSS graceful degredationfor this.
- Audio? Web Audio in order of preference – OpenAL, Audio Data, RIA Audio HTML5 audio.
- AJAX? CORS is best, followed by a proxy, JSONP, then a hybrid app.
Up now, the Great HTML5 Divide by Ray Bango of Microsoft. Works on jQuery, HTML evangelist. HTML5 is great, it’s definitely the future, but there is a divide. Talking about browser fragmentation today, feature detection, polyfills and shims. How do you leverage HTML5 and CSS today? Solutions will be provided.
Biggest problem now is non-modern browsers. caniuse.comis a good resource for figuring out which browsers support. Lot of browser fragmentation, across browsers, within the same browser, different versions.
Feature detection is a good thing – do it. Don’t do browser-based detection.
Modernizris great for feature detection – detects all major features for CSS3, HTML5, etc.
Polyfills and shims – polyfills are used to fill in cracks – shims fill in cracks. SVG support, Web Storage, WebSQL, WebSockets, etc. Consider 3rd party library– developers may not want to support it anymore.
Take-aways: avoid browser detection, use polyfills, do smart fallbacks.
Christopher Slye (Adobe Systems, Inc.): Shortcuts: Web Typography
Next up, Christopher Slye from Adobe to talk about Web Typography. Involved in WOFF standards group. Two real font formats – TrueType and PostScript. Going forward, quality will be about the same. EOT is Microsoft’s font container – been exclusively for Internet Explorer. WOFF is new Web font format – will be the first interoperable Web Font Format.
WOFF is a compressed font container, has font metadata. CSS3 Fonts Module – contains entries for properties size, weight, style. OpenType has nothing to do with the Web. Kerning, ligatures and alternates.
CSS font spec allows you to “optimize readability”, which uses ligatures. Discretionary layout features. Good typography on the Web requires font designers to be able to specify font choices on a per-character, per-stroke basis – let’s typographers add case shifting, kerning, alternates, small caps, oldstyle features, etc. – all help readability.
Next up, Paul Irish from Google talking about The Foundation of the Web Platform – HTML5. Talking about HTML5 underpinnings – on the jQuery team.
Explains that came about because browsers don’t really care about all of the variations and that all of the variations are difficult to understand for Web developers. was introduced because of a security vulnerability in UTF-7 and script tags.
Browsers used to parse broken markup in a different way. Now they all parse broken markup in the same way. There are optional tags – html, head, body, etc.
You can not use quotes in HTML5 attribute IDs, you can use many UTF-8 characters (like hearts and the unicode value for a snowman).
The HTML5 spec is a repository of valuable browser knowledge – acquired through copious reverse engineering.
W3Conf LiveBlog – Day Two [Manu Sporny, Digital Bazaar, Nov 16, 2011]
Giorgio Sardo (Microsoft): HTML5 Demo Fest: the Best from the Web
Some amazing SVG demos happening on screen at W3Conf. An SVG music video, amazing filter effects with SVG, CSS applied to SVG, etc.
Amazing HTML5-based demos. BMW using canvas to preview cars interactively using open standards. The Killers (the band) launches a pure HTML5 and canvas page. Amazing TRON interactive comic book done in pure HTML5.
CSS3 demos showing off animation. Amazing font work showing off WOFF (new font stuff in HTML5). Associated Press showing interactive news demo showing top stories in an interactive way, using Local Storage to save stories for offline viewing.
Demos showing that pen-based input, touch inputs, motion inputs on HTML5 pages – no plugins necessary. Showing smooth transitions between HTML5 video and page content, giving a very smooth website experience.
Current presenter is Giorgio Sardo, doing a great job showing off the power of HTML5 today. Showing off the need for multi-touch events – talking about mobile phones or large screens shared among different people, running HTML5 Apps.
Moving on to WebApps and the File API. Showing drag-drop of files from the desktop to the browser window. Drag-Drop API allows you to take files from the desktop and drop them into a web page – file information and the byte stream is made available to the browser. Web Performance API allows you to know when someone isn’t viewing a page and allows you to shut down processor intensive or battery draining portions of Web Apps.
HTML5 games being demoed now – Angry Birds in pure HTML5, 65,000 PacMan mazes in pure canvas, high speed, smooth interoperation.
See @gisardo on twitter for a link to all of the amazing HTML5 demos.
Mike Bostock (Square): Shortcuts: Data Visualization with Web Standards
Mike Bostock is up now, talking about Data Visualization with Web Standards. We should not put a focus on charts and graphs, but rather visualizing information in a way that is closer matched to the data we’re trying to understand. D3.js – Data Driven Documents.
D3.js looks at data visualization as a document. Map a quantity/value to a symbol then render. decoupling data from visualization. D3 mixes SVG, CSS and HTML togetherto build beautiful visualizations – for example hierarchical data sets can be visualized as hierarchical bar charts, or bubble graphics, and smooth transitions between them.
Showing off how D3 also allows you to do interactive visualization – select part of the data set, see highlights elsewhere in the data set. The tools we create don’t exist in isolation – combining SVG + Canvas + CSS allow us to create very compelling visualizations.
Becky Gibson (IBM): Making Accessibility Mainstream
Becky Gibson, from IBM – Senior Technical Staff Member, on stage now.
Issues with Accessibility – Vision issues, Mobility issues (not everyone is able to use a mouse), temporary disability, learning disabilities, hearing disabilities, age. Eventually all of us age – every one of us will have accessibility issues.
Why care about Accessibility? Government/legal reasons, monetary reasons (1.75B of disposable income from people with accessability issues), independence, altruism – it’s the right thing to do. People with disabilities don’t want a hand out, they want your respect.
Enabling Accessibility. Use semantic HTML, use alt-text on images, label form elements, add keyboard support, use tabindexes, support high contrast mode (all background images are removed).
Assistive Technology Support – WAI-ARIA – uses role attribute and aria-* attributes to enable accessible applications. Multiple roles – tree, spinner, grid, many more. States and properties – required, expanded, checked, owns, many more.
ARIA Roles landmark the main roles on the page – banner, main, navigation, search, form. Popular screen readers support landmarks. ARIA Presentation role – preferred way to say that something is being used to present some information?
ARIA states and properties. ARIA allows one to specify that a region of the page has been updated and how one should be notified that it’s been updated. For example: “aria-live=”assertive” aria-atomic=”true”. aria-required indicates that a field is required. aria-invalid tells us that something someone has entered is invalid.
The future – mobile – VoiceOver screen reader and ARIA support is in Safari – iOS has the accessibility advantage right now. Intent-based Events – how do you do generic events?
Brad Hill (PayPal), Scott Stender (iSEC Partners): Securing the Next Generation of Web Apps
Up on stage now – Brad Hill from PayPal and Scott Stender (iSEC Partners) talking about Web App Security. We need a way to protect information from prying eyes on the Web. New threats are client-based – Cross-Site Scripting and Cross-Site Request Forgery.
Old clever tricks – like script tags, iframe elements, JSONP are mature, but you have to understand what security model you’re signing up for– you’re effectively giving the source of a JSONP data/iframe check-in rights to the code on your page. So, if you use these clever hacks, be careful how you do it and who you do it with.
New features coming up that help web security – CORS, XHR2, WebSockets – powerful tools. Auditing is more difficult with these new technologies. postMessage (communication between two tabs in the same browser) and Web RTC (real-time data/media streams). These technologies, while very cool, open up the attack surface on the Web App.
Assets and attack surface are moving to the client, but the focus is still on securing the servers. Future of Web App Security is in the client, not the server. We are falling behind on making sure we’re securing our Web Apps.
What can we do to mitigate these threats? Comprehensive testing/verification, do tests on production code. Get rid of built-in “Game Over” security threats – JSONP is bad, plaintext HTTP is bad. Create code that is designed to be securable– compartmentalize the code, decouple the code, be explicit, do test-driven design.
More specific thingsyou can do: Use HTML5+ Mashup APIs, use HTTPS (and authenticate your origins), secure design – compartmentalize and sandbox origins, do good client-side testing.
For compartmentalization: use credential state isolated from the DOM, minimize foreign origins, create unique origins to isolate apps and sandbox dangerous or active content.
You can sandbox bad JSONP code using postMessage.
Implementation validation is also important. Test using security tools – DOMinator, DOM Snitch, WebDriver and Selenium for testing. Start using them, contribute bugs, make the tools better.
Content Security Policy – pioneering work at Mozilla/Google– Least-Privilege environment. Let’s you say: “No in-line script, no code from strings via eval(), no data URIs, code must come from libraries with origins specified in a whitelist, origin whitelist for images, media, frames, fonts, plugins, etc.”
Click-jacking and UI Redressing is still a problem with Mashups. DOM is still a mess – lots of browser-specific quirks.
Grant Goodale (Massively Fun): Shortcuts: Touch Events
Grant Goodale up now from Massively Fun (@ggoodale) to talk about touch events. First touch API appeared in iOS 2.0, now a part of the HTML5 spec.
Many differences still exist between vendors – many of the browsers don’t support it yet, but there are lots of mobile browsers that do support touch events. Many different types of input, you want native-like behavior. Firefox supports touch events.
Cross-platform multi-touch webapp checklist – disable standard gesture handling, handle rendering in a loop outside of the touch event handler, support single-touch devices (like Android phones), handle mouse events (phones w/ a physical pointer, phones with no touch events, desktop browsers).
Grant Skinner (gskinner.com): Hello. Games. HTML5 Gaming Today.
… by Grant Skinner (gskinner.com) – creator of Eazel.jsand Pirates Love Daisies.
Traditionally worked with Flash – moved to working with HTML5. Games on the Open Web today and in the future.
What do you need to make a great game? An idea – you can be inspired by the technology you have available to you. Can we design games designed for the Web? Loading – critical piece, how do you get the game onto the client? Monitoring progress is difficult in games (no clear standards to see how long it’ll take to load everything).
XHR2 is great because it has progress events – ability to load in binary assets. caching – minimizes bandwidth/connections – cache manifest is good, but it’s an all-or-nothing solution. If you change the cache manifest, everything downloads again. HTML5 local storage is good, tricky to store non-text assets, low-level way of supporting cached data – lots of custom implementation work.
Graphics – lots of options. DOM and SVG is good – interaction model is good, ubiquitous, hardware accelerated… but high overhead (can be a performance killer). Hardware acceleration works really well for Canvas, not so great for SVG and the HTML DOM. WebGL (based on OpenGL ES2) – fast, low-level hardware graphics, 3D vertex shaders, extremely sharp learning curve – Three.jsmay help, but still very difficult to use. Canvas 2D is broadly available, including mobile, consistently implemented, easy to get started with, increasingly performant.
Combining surfaces is good– combine Canvas and SVG and DOM. HTML DOM does the UI very well – all UI in their games is done with HTML – use Canvas/SVG for graphics. Using pluggable renderers – you can switch between three.js, EaselJS based on your platform – write the content once, pick the renderer based on the device.
The Web Audio API is coming– very powerful API for real-time audio manipulation, runs at native speed, synthesis, analysis, mixing – limited support now, but will solve a ton of problems.
Basic interaction – keyboard capture is incomplete, but mostly adequate. Mouse interaction is fine for point and click games. WebGL and Canvas require custom handling of mouse events. Full-screen and interaction API is a bit quirky. mouse lock and synthetic mouse events are necessary when you want the cursor to go off-screen and still generate mouse events (which you need to do in first-person games when walking, running, navigating).
Upcoming device APIs – touch events, orientation change API, device orientation API, nothing for orientation lock yet (don’t change the orientation when playing a game).
Communication – XHR (http requests and polling), WebSockets (text only for now, but binary coming soon). WebRTC / Peer Connection API – peer-to-peer audio, video and data. Network Status API – is the device online or offline.
Asset preparation (for designers) 3D – Inka3D exports to WebGL. Sprite sheets – Flash, Zoe, TexturePacker. 2D – Wallaby, SVG to Canvas, EaselJS.
Marketing and Monetization – Web games are great because you’re already in a social medium.
Flash is more ubiquitous, it has more tooling, building a game is cheaper in Flash… but it doesn’t run in iOS, probably won’t run in mobiles, Windows 8 Metro won’t run it either.
Exciting times for Games in HTML5 – still some rough edges, but technology is maturing quickly.
Faruk Ateş : Shortcuts: Modernizr
Design for the browser of the future by detecting features instead of browser versions – useful for progressive enhancement, graceful degradation, regressive enhancement.
Modernizr allows you to test a feature and then conditionally load code that expects the feature or code that uses a polyfill.
People arrive at your site for the content, but will leave very quickly if the site isn’t fast. Showing demo of progressive enhancement.
How to use modernizr? Go to http://www.modernizr.com
Development and production version. Good for CDN distribution.
Art Barstow (Nokia), Paul Cotton (Microsoft), Tantek Çelik (Mozilla), Charles McCathieNevile (Opera), Chris Wilson (Google), Peter Vosshall (Amazon):
PANEL: Browsers and Standards: Where the Rubber Hits the Road
What’s interesting about the browsers?
- Nokia has shipped millions of browser installs – written from the ground-up in some cases – shipped browsers from Opera – also shipping Windows Phone browsers – large swath.
- Microsoft – IE9 and IE10 – driving home the message about all web developers using the Web Platform – picking stable specs, moving it into the browser.
- Mozilla – we’re a non-profit, our focus is a bit different – mission driven focus– we’re concerned about SOPA and censorship over profits.
- Opera – we’re a commercial company, we are out of Norway, we are product focusedfor our customers.
- Google – focus on making the Web platform better.
- Amazon – just got into the browser game, built from the ground-up to use Amazon web services infrastructure – browser running on a mobile device.
What do you think about DART and SPDY and how they were brought to market?
- Paul Cotton (Microsoft) – innovation is good, we should have more of it but we need to coordinateon those things.
- Paul (Amazon) SPDY is great, we ship it, it reflects what’s happening in the real world today.
- Chaals (Opera) browser vendors are just one part of the puzzle, developers need to take part in it, security folks need to take part in it.
- Tantek (Mozilla) – the problem is delayed open – for a fair look on the problem search for “delayed open google microsoft”.
[ 20 Nov: @t Tantek Çelik also @
#W3Conf: fight “delayed open” tactics per Eran’s post: Open vs. Fast, Good vs. Evil, Google vs. Facebook: ttk.me/t4Ex4]
Vendor prefixes, are they outmoded?
- Chris Wilson (Google), vendor prefixes are not that good.
- Chaals (Opera) bad authoring is worse, so is bad Web teaching, we need a mechanism for innovation – vendor prefixes are terrible, except for all of the other things we tried, which are worse – sunset the vendor prefixes.
- Paul Cotton (Microsoft) we want vendor prefixes to be taken out when we go to Candidate Recommendation status – we need to get to CR faster, we need to do our specs in a much more modular fashion. Part of the standards process is at fault here – standards need to move faster.
- Tantek (Mozilla) The better that we can get about dropping vendor prefixes, the better. Vendor prefixes suck, but it’s the best we’ve come up with so far– we could do better – we’re all open to suggestions on this point.
- Paul (Amazon) Vendor prefixes are a form of technical debt, you have to pay it down eventually.
Tantek asks why VIDEO element didn’t have a vendor prefix – but it still worked… why?
- Chaals: It was easy to understand what Video should do.
- John Allsopp: It had a good fallback.
- Tantek: Video wasn’t easy, but we got it right – why was that? Arguing on stage
What about operating system integration, like with Chromebook?
- Tantek: it’s bad – robs the user of choice. iPhone, no choice – Chromebook – no choice.
- Chris (Google) – Chromebook is just the Web layer, you can replace the OS if you want to– there is choice there.
- Paul (Amazon) having HTML5, creating a true application environment – maybe the model of using the OS as the browser is terrible.
How do most of the regular web developers know what should be used and what shouldn’t?
- Chaals (Opera) All of us building the Web need to go out and learn and teach other people. There is nobody spoon feeding information to you – this field changes so quickly – look around to your peers, they will help you.
- Paul (Microsoft) Browser vendors need to hang out with one another – we need to work together, get things done. Community Groups at W3C are going to be key – get groups going with least amount of overhead, understand what minimum required to get a spec done is needed. The way we’re dealing with the HTML5 spec is bad – it’s a flawed way of working on a standard. We will continue to work together at W3C. We have a real obligation to the community here.
- Tantek (Mozilla) – it’s not sufficient for browser vendors to work together – we need to work together IN THE OPEN.
- Chris (Google) One of the goals behind DARTwas to provide something simpler.
- Chaals (Opera) it would be terrible if the solution was not interoperable.
- Chaals (Opera) different societies define privacy differently, difficult to start the work when you don’t know where you’re going.
- Chris (Google) people should have the tools to control their own privacy, you can’t just leave it to the end user.
- Doug (W3C) Good work on do not track, cryptography going on at W3C.
What about internationalization?
- Paul (Microsoft) We need to be able to ship browsers worldwide – how to bring rest of the world into the conversation.
- Tantek (Mozilla) There are 70 versions of localized Firefox browsers – built by the international community. Customized for specific locales and markets.
Women in browsers, why are they not represented?
- Paul (Microsoft) I work with 2 women on my team, 3 men. More women attending at TPAC, this is a good sign.
- Chaals (Opera) different societies have different levels of participation.
- Paul (Microsoft) Women are certainly welcome here – we want much broader participation than we have now. I think some of the problem is societal, it’s not a social norm in the USA – we need to get better about teaching science and math to he smarter of the genders.