Drupal 8 HTML5 Initiative Meeting #6

We encourage users to post events happening in the community to the community events group on https://www.drupal.org.
Jacine's picture
Start: 
2011-08-09 16:00 - 17:00 America/New_York
Organizers: 
Event type: 
Online meeting (eg. IRC meeting)

Agenda

Please join us for our 6th bi-weekly meeting to discuss issues and progress related to the Drupal 8 HTML5 initiative. The meeting will be held in #drupal-html5 in IRC at 4 PM EST. See time zone information. During the meeting, we will discuss the progress of the current sprint and attempt to resolve any issues/roadblocks holding them up. We'll also tag issues for the next sprint: HTML5 Sprint: August 2011 - 2. Please post your proposed discussion topics in the comments if you have any. You can subscribe to the Google Calendar for this and future meetings via the iCal or XML feeds.

Transcript

See attachment below.

AttachmentSize
drupal-html5.8-9-11.transcript.txt47.58 KB

Comments

We had a great meeting today!

Jacine's picture

We had a great meeting today! Thanks to everyone that attended and participated. :) The chat transcript has been attached to the original post (above). Here's a summary of what we discussed:

IE7

drupal a11y's picture

I hope until D8 is out the IE7 usage is also nearly zero. But Modernizr is a good solution.

According to http://www.w3schools.com/browsers/browsers_explorer.asp

2011 Total IE 9 IE 8 IE 7 IE 6
June 23.2 % 3.6 % 12.9 % 4.4 % 2.3 %

2010 Total IE 9 IE 8 IE 7 IE 6
December 27.5 % 0.5 % 16.5 % 6.1 % 4.4 %

But this may vary, especially in business environments.

The Wikipediapage has very different numbers from market share:

Internet Explorer Market Share
— Jul 2011[82] via Net Applications [note 1]

Internet Explorer 6 9.22%
Internet Explorer 7 6.25%
Internet Explorer 8 29.23%
Internet Explorer 9 6.80%
All variants 52.71%

Not to be a dick, but

seutje's picture

Not to be a dick, but w3schools is like the worst reference on the entire intartubes, for anything, including HTML and JavaScript. http://gs.statcounter.com/ is a much better reference ...
also see http://w3fools.com :P

We need to start thinking of

LewisNyman's picture

We need to start thinking of ways to dynamically load JS files based on rules such as support for CSS selectors or html5 elements.

An asynchronous script loader like <ahref="http://yepnopejs.com/">yepnope.js is a good option. Apparently loading the scripts in this fashion is faster then concatenation cached article for most sites.

Discussion underway

rupl's picture

We have already started thinking about this, and the transcript is available.

Three possibilities were brought up: Modernizr custom build, yepnope, and Selectivizr. Right now the feedback from core Drupal maintainers is that full Modernizr is too heavy (and MIT licensed, although that may not be a problem). We did not yet hash out which components of Modernizr we'd like to include if we went that route. Similarly, we did not thoroughly discuss yepnope.js or selectivizr yet.

For older browsers that don't support HTML5, here's the HTML5shiv in core issue, which for the time being would not be loaded asynchronously. Additionally, it's being worked on in a jQuery issue, and by the time D8 comes out the shim/shiv may be included in jQuery core, requiring no further action on our part.

What I've posted here is a summary of that transcript above. If we want to keep discussing we should probably move it to the issue queue.

Thanks, I had a quick skim

LewisNyman's picture

Thanks, I had a quick skim through the transcript.

I've created an issue here: http://drupal.org/node/1252178

For the record I don't see why MIT is an issue. MIT allows you to re licence, if we wanted to we could publish Modernizr as GPL.

Instead of just including the Moderizr JS file in core, why don't we build the Download Builder right into core? We can allow modules and themes to call Moderizr through an API to include certain tests or functions.

This is an old debate

rupl's picture

Yes, the Drupal Licencing FAQ says GPL-compatible is ok, but in practice, it has not been observed. Yes, it's common knowledge that you can relicense MIT, but now you've created a fork that you have to maintain manually. Additionally, even though it seems legally kosher, the Drupal Infrastructure team has business rules that disallow including such code.

In the chat transcript above, Crell (member of infra team) mentioned that MIT code is closer to being considered due to WSCCI. For now let's drop this conversation because it has already happened many times, even recently.

All this being said, I'm not against the idea.

Biggest problem with

seutje's picture

Biggest problem with shim/shiv in jQuery core afaik is that it would force u to load jQuery in the head, which is... well, nasty!

Modernizr?

awats008's picture

This has been brought up already but wanted to mention Modernizr has yepnope.js bundled (Modernizr.load()) and additionally handles shims/ polyfills and feature detection.