Open Web Platform Weekly Summary - 2011-11-21 - 2011-11-28
This week, one of the main discussions has been around developing (or not) a support for XPath in find
and findAll
methods. The Open Web Platform weekly summary is also mentioning Web architecture, Web Apps WG hosting new work.
HTML5
The old HTML4 abbr
attribute has been deprecated in the HTML5 specification. The role of the attribute was to give a short form of table cells content. It was meant to help users getting the content of these cells quickly. A Firefox patch has just been proposed to implement it.
Frank Olivier (Microsoft) said:
Text editing is certainly a fool’s errand in canvas.
and indeed there was previous attempt to recreate text-editor all in Canvas. Some of these projects have been abandoned since. That said the group is struggling to find solutions for raising accessibility in canvas to an acceptable level. One solution which is being explored is to add primitives for Path.
Web Apps
There are proposals for
- a community group working on best practices and developer guidelines for building apps and webapps that use Web technologies in a “network-friendly” way
- one on the concept of Native Web Apps. New work on W3C widgets will take place there as announced by Marcos Caceres.
The work on Component Models that I mentioned a few times about under the label shadow DOM is moving to the Web Apps WG led by Dimitry Glazkov (Chromium team).
Darin Fischer (Google) proposes to add the Pointer Lock (formerly known as Mouse Lock) spec and the Gamepad spec be added to the Web Applications WG’s charter.
ECMAScript
Some fundamental features are missing in JavaScript. I know for example missing things like startsWith
and endsWidth
on strings annoy me a lot. There is a proposal for evolving ECMAScript on IE blog.
DOM
There has been a gigantic thread (with a lot of misunderstandings and rebuttals) about allowing XPath in the find
/findAll
APIs we were talking about last week. The discussion goes along the common permathread about CSS selectors and XPath has a way to select a path in a DOM. With similar goals, they often addressed different problem spaces and they do not have the exact same set of features. Some people argue it is not worth the cost adding XPath for selecting nodes. Eventually, people will reach an agreement. We are not there yet.
matchesSelector
is verbose and people start to look at ways to make it shorter for Web developers. Two proposals have been by Tab Atkins (Google) with .matches()
and .is()
. Though Dimitri Glazkov said that he wishes to use .is()
for components. It would be used like
elt.matches("div span")
CSS
Jake Archibald (Lanyrd) is not satisfied with the Shadow DOM and scope stylesheets we mentioned a few times in that column.
Experimenting with new styles for CSS Specifications. CSS shaders is currently having the proposal.
Web Architecture
URIs are one of the corner stones of the Web architecture. There is a specification clearly defining the URI syntax and meaning. But as usual with the human Web, things get deployed with errors in a distributed way. What is happening when you get something which looks like an URI but is not really a URI. User agents have for long implemented techniques to cope with the common URI mispellings found on the Web. Mike Smith will start working on a document on how browsers process URIs, following a proposal made at HTML WG F2F during the TPAC 2011. It has been suggested that this should be part of the URL API document.
HTTP
The Referer
HTTP header has been a concern for a long time in terms of security and privacy. Adam Barth is proposing to add a referrer
attribute in HTML (meta
element) for suppressing its value from each HTTP requests.
This week, the theme of Anne Van Kesteren's report is about XMLHttpRequest.
This column is written by Karl Dubost, working in the Developer Relations team at Opera Software.
Hi Karl,
In regards to firefox implementation of the abbr attribute its more than 'proposed', its inbound "Verified fixed in Mozilla/5.0 (Windows NT 6.1; WOW64; rv:11.0a1) Gecko/20111127 Firefox/11.0a1".
I would suggest the continued highlighting of the 'text editor' bogeyman, does not serve to find solutions for the real world canvas accessibility issues yet to be resolved, but it interesting to note the work on JavaScript based IMEs. Also Charles Pritchard has provided a summary of discussions on canvas accessibility, which is worth a read [public-canvas-api] Summary of accessibility threads between 2009 - 2011.
Ah indeed. Thanks for the clarification Steve. That is very useful.
I'm sorry to be not too close to the HTML5 discussions, but what on earth is the rationale behind dropping the abbr attribute?
It is useful and we've used it a lot in my company for a long time.
Where is the cowpath and why are we putting stones on it so that cows must walk elsewhere?
This plus longdesc, which is a long-standing discussion point (and I still haven't seen a proper, solid replacement for it), makes me think my own implementation of HTML5 will not be standard. But at least it will work for blind people.
Stéphane, not sure you have seen Steve's comment here. Firefox has implemented it recently, and if others are following, I guess it is likely to be reopened. I'm pretty sure if there are implementations experience of abbr in other products that would be valuable input for people.
Karl,
Sorry for being late responding.
The abbr attribute for td has been implemented for a long time in Jaws when one reads tables, and is very useful to us on intranets.
(and sorry for not being able to contribute more to the discussions, but I feel, like Steve, that 'paving the cowpath' and 'provide us with figures' only works one way most of the time - sorry if I sound bitter)