W3C Blog

I have decided to change a bit the style of weekly summary of the Open Web Platform. Instead of just going through the list of mails, I will try to focus on more specific things and give more context about them. Anne Van Kesteren published his report on August 24.

• “How to organize a W3C Specification” is a common question at W3C. This time it started about the DOM specifications. The argument falls both on side of the consistency, work style, and technology design. There is no ultimate and definitive answer. In my opinion, you need to choose the format that will maximize participation for your context. The specification organization might be also influenced by the way you want to design the technology.

• A new group for the Web Education/Evangelization community has been opened. Chris Mills is the co-chair of the group with Doug Scheppers. People on the mailing-list have started to introduce themselves. Join! It is free.

• A while ago, I have talked about the Shadow DOM. The discussion evolved led by Dimitri Glazkov (Google) and Dominic Cooney (Google) into the Component Model, a way to introduce support for creating custom DOM elements enabling layout managers or author UI elements included in the page, but not really part of the page. It inherits from the work done on XBL, XBL2 and MS HTML components, but will be more modular. See also Component Model Update mail and also this one.

• Your days of DukeNukem are not over. There is a proposal floating for a Joystick API. With the rise of HTML5 games popularity, there is a growing interest for this proposal.

• registerProtocolHandler gives the possibility to Web sites to become handlers for specific protocols. It is currently implemented in Chrome 13. There was a proposal on the list to extend the protocol.

• Would it be nice to be able to define how to proceed with undo/redo features but inside the browser? Ryosuke Niwa (Google) has proposed a specification along use cases. Ian Hickson decided to remove the old UndoManager so that people could focus on Ryosuke Niwa’s proposal.

• MouseLock is a proposed API to give the possibility to block the mouse in some circumstances. Vincent Scheib (Google) covered some of the issues identified with the proposal.

• Digital images contains a lot of hidden metadata such as date, focal length, etc. These are contained in the EXIF part of the image. Some camera includes the orientation of the image when taken. Tab Atkins (Google) is proposing to add a naturalOrientation property to the img element so it could be used in-browser photo-editing tools.

• Be careful with the CSS Flexbox (a feature giving you the possibility to reorder the boxes in CSS). This is one of the CSS module which has been quickly implemented by some browsers vendors. Too quickly? Web developers have started to use it and there is already some legacy code online. Why legacy? Flexbox is going through major revisions with backward incompatible changes. Tab Atkins (Google) is working right now on the layout algorithm.

• The Web is evolving from a Web of documents to a Web of applications. This creates new constraints and usages to the system. HTML5 created history.pushState() and history.replaceState() for managing the history of a navigation and gives proper and usable URIs. Ashok Malhotra (Oracle) published a new version of a W3C TAG finding on Identifying Application State.

• There is a request on allowing @autofocus for any elements having @tabindex property in HTML. This would give the possibility to get a frame like behavior for some blocks in the document. When people create non scrollable section in a Web page, it creates a load of secondary issues.

• Channy Yunn has proposed a Web Crypto API Community Group. It has not been accepted yet. If you are interested, join the effort. If you need more information about the topic, see Web Crypto API and DOM Crypto API

• Aryeh Gregor (Google) has proposed to extend the DOM for allowing authors to specify modifiers for accesskeys. Right now, it is not possible to reassign the modifier which is dependent on the system (for example, the command on macosx and Ctrl on windows).

• When dealing with a document, you might want to change the name of an element in the DOM by another name. Aryeh Gregor (Google) has defined in the editing specification an algorithm for it. The discussion revived an existing feature called renameNode. The current property is still missing features, but there is a will to include it in DOMCore.

This column is written by Karl Dubost, working in the Developer Relations & Tools at Opera Software.