HTML5 Wish List for 2011: Interview with Michael Mullany
Michael Mullany from Sencha has published a list of things that would benefit HTML5 during 2011. InfoQ has interviewed Michael in order to get some more details regarding his vision.
Michael’s original list includes the following:
- Sustained effort to move more -webkit effects into CSS3
- CSS3: A Richer Effects Toolbox
- High Performance position:fixed for mobile
- GPU Acceleration
- Deeper Device Access
- Better debugging tools for mobile browsers
- Web sockets stabilization
- IE9 With Complete CSS3 Support
- A HTML5 codec armistice
- A Reboot for WebSQL standardization
InfoQ: Could you give us a list of the most important effects that you feel should make it into the CSS3 spec? What kind of new functionality will these effects provide to developers?
Michael: Getting what's in the modules implemented cross-browser is by far the most important thing that can get done :-) But the top wishes we have are to move more of the WebKit-specific effects that are not yet on standards track, into CSS drafts. This includes advanced text styling, masks and font effects.
InfoQ: With IE's slow adoption of new Web standards and technologies, when do you think these effects will make sense to desktop developers?
Michael: I think the more that IE drags its feet on functionality, the more that even the most conservative enterprises will begin to look to Chrome and Firefox for modern browser functionality. We think that the frameworks (like Sencha Ext JS) will be in the forefront of adoption because they can create IE-compatible implementations using IE-specific technologies like VML and filters, without the developer having to learn the IE technology stack.
InfoQ: You mention that currently you need a framework like Sencha Touch in order to have fixed position UI elements. Why is that?
Michael: Mobile WebKit doesn't support fixed-position UI elements; the CSS to support this has been turned off in the browser. This is due to the browser being optimized for pan/zoom of web pages designed for desktop-size screens.
InfoQ: You mention that you'd like better debugging tools and Senche has already delivered a solution for remote script debugging on Android devices. What other kinds of tools should we be expecting in the near future for HTML5 specific functionality like canvas, animation, Websockets, etc.
Michael: Lots of debugging tools exist for desktop, so it's really a matter of getting access to the mobile browser internals rather than creating tools specifically.
InfoQ: What are the issues with Websocket regarding adoption and the spec itself?
Michael: The issues are that browsers have been shipping incrementally updated versions of the websockets protocol while the spec itself has been evolving. For example, Chrome 4 shipped with a very early version of websockets that is now outdated (luckily, almost no one is on Chrome 4 any more). The iPad originally shipped with a websocket implementation, then it was removed, and then it was added back. Firefox 4 beta also originally shipped with a websockets implementation, which was then removed as a result of this security issue: recently, it was discovered that a very small percentage of proxy servers (<5%) do not correctly handle a HTTP mechanism that websockets uses. This introduces the possibility of those servers being used to serve an attacker's payload. While this is arguably a proxy server bug, the authors of the paper are advocating that the websocket should no longer use this mechanism and instead use an alternative.
InfoQ: It seems that your wish regarding the video codecs is getting closer to happen as Google announced they are dropping H.264. On the other hand, with H.264 available and royalty-free for free internet content, why would content providers care to pay for the transition to WebM? If you deliver paid content, don't you need digital rights management, advance streaming options, etc., which are only available in Flash?
Michael: H.264 is free to distribute video but encoding and transcoding is not free; that still requires the payment of licensing fees. Both Flash and Silverlight offer more advanced DRM and streaming video options (I believe the Netflix streaming client for Xbox is Silverlight-based and offers dynamic adjustment of streaming rate without rebuffering). The point is that we need an unencumbered codec for everyone who's not interested in DRM.
InfoQ: What do you think was the rationale for W3C to drop the support for the relational browser storage (WebSQL) in favor for a hierarchical one (IndexedDB)? Why do you propose that the WebSQL standard should be made active again?
Michael: The rationale was clear: in order for a technology to make it to a standard, there needs to be two independent implementations that are built from the spec. The central concept of a complete spec is that anyone can take the spec and write a browser that works with any content that conforms to spec. In the case of WebSQL, all the implementations are simply SQLite embedded in the browser: no one was willing to write a new SQL implementation from scratch. In addition, there seemed to be a feeling from the Mozilla camp that SQL was a heavyweight, inelegant query mechanism that we shouldn't be keeping alive.
InfoQ: What other wishes do you have for HTML5 in 2011?
Michael: I think a faster Canvas would be high up there on the list.
For more HTML5 insight you can join us in QCon London, where there will be a whole track focusing on “HTML5, the Platform”!