Please leave your sense of logic at the door, thanks!

Archive for June, 2011

WHATWG Weekly: Event Constructors, Augmented Reality, and a new HTML5 Timeline

Monday, June 27th, 2011

Still behind with most of the specifications I edit (as ever, really), but did catch up with what went down while I was away and tracked what happened during the week.

Changes to the specification

Errors in workers (the best place to write synchronous code) now propagate up all the way to window.onerror. In addition, self.location in workers now stringifies.

The postMessage() API became a little more complicated for forthcoming support of transferring ArrayBuffer objects. Instead of cloning the ArrayBuffer object this means that the sender ceases control of it. As cloning is no longer required using this functionality would be a performance win for applications. (And less memory usage.)

Mailing list

On the mailing list details of collapsed whitespace handling for the contenteditable attribute are hashed out, as well as getSelection() in the context of vertical writing modes, HTTP adaptive streaming of video, drawing with singular transforms and zero-sized gradients on canvas, and a few other things. There were also some requests:


The DOM Core specification gained support for event constructors. Once this is shipped in browsers you will be able to do var ev = new Event("yay"). You can set attributes via a second argument: var ev = new ProgressEvent("progress", {loaded:42}).

In W3C bug 10623 the future of exceptions on the web platform is being discussed. Whether we should continue using DOMException or a new approach that relies less on the awkward code attribute checking. Input appreciated!


Maciej posted a revised timeline for the W3C’s work on HTML5 (WHATWG develops a parallel edition of HTML that continuously evolves and has no version number). The W3C HTML5 draft is in Last Call which means your technical comments have to be filed by a certain date. That date is 3 August 2011. From that date onwards it is expected to take until new years until all feedback is addressed by the editors, and until end of April 2012 until the Working Group has verified. (At the moment the HTML WG charter calls for the HTML5 Candidate Recommendation draft to be published at that time. It seems likely however another Last Call draft is required first. Feedback has already resulted in substantive changes.)

Earlier Maciej also posted on editorial assistants that will help Ian out with dealing with the Last Call feedback.

Posted in Weekly Review | Comments Off on WHATWG Weekly: Event Constructors, Augmented Reality, and a new HTML5 Timeline

WHATWG Weekly: Back!

Monday, June 20th, 2011

Big round of applause for Shelley for taking over in my absence. Much appreciated. As I am still somewhat catching up myself with everything that has happened in the last three months I decided to summarize changes to the WHATWG specification since May 23 and highlight a few other efforts I noticed. Hopefully next week I am back up to speed.

Changes to the specification

The send() method of the WebSocket object now accepts ArrayBuffer and Blob as arguments. I.e. the Web sockets API supports binary data. On the receiving end you can specify the type of object you want using the new binaryType attribute.

The application cache support gained support for caching of cross-origin resources over a secure HTTP connection.

Parsing of elements related to ruby markup was made more forward compatible.

A new typemustmatch attribute was introduced for object to ensure only specified plugins can be initialized. If you link cross-origin resources from object you should use this feature to prevent your users from being attacked.

The EventSource object is now enabled with cross-origin support, making it possible for developers to use different servers for the streaming aspects of their application.

These were just the highlights. Numerous other changes have been made, e.g. the WebVTT file signature was simplified, a security check for the AddSearchProvider() API was removed, the error and close events were removed from PeerConnection, et cetera. You can use the HTML5 Tracker to find out more.


A little before May 23, but still worth highlighting. Adam Barth tackling processing of URLs in Parsing URLs for Fun and Profit. URLs are of course fundamental to the web, but much like any other piece of technology deployed on the web (e.g. HTML, CSS, HTTP) URLs have interoperability issues. As with the work done by WHATWG on HTML, the plan is to take those away.

Alexey Melnikov and Julian Reschke meanwhile are drafting Update to MIME regarding Charset Parameter Handling in Textual Media Types attempting to finally fix the brokenness of the ancient RFC 2046. In particular, how encodings are handled for text/* media types.

Posted in Weekly Review | 2 Comments »