Drop legacy browser support

[necolas]

IE6 is essentially dead in the Americas and Europe...and increasingly rare in the territories where it is clinging on. IE7 is close to dropping under 2%. Firefox 3.6 is no longer supported by Mozilla. Safari 4 died ages ago.

Going forward, we shouldn't actively support these browsers in the HTML5 Boilerplate project. If in the future someone absolutely needs legacy support for a specific project they are working on, then they can download the last stable release of HTML5 Boilerplate that supports it - it will do the job perfectly well.

Probable changes involved in dropping IE 6/7, Firefox 3.6, Safari 4:

• Changes to conditional comments around opening html tag.
• Switch to normalize.css v2
• Update main.css to remove any IE 6/7 hacks
• Change version of IE that gets Chrome Frame prompt.

[alilleybrinker]

Sounds good to me. My only concern is that, based on the IE6 countdown page, IE6 still has a 23.8% browser share in China. While not a majority, it is about 1/5 to 1/4 of the total market in country.

Obviously, higher share in an individual country isn't a good enough reason to maintain support for an outdated browser. However, it may just be worthwhile to include something in the Wiki that specifies the relative importance of IE6 to users in China.

[jonathantneal]

I'm ready for an IE8+ world. IE8 can be the new IE6.

[paulirish]

I'm ready for an IE9+ world. :)

But... more realistically..

Big 👍 on dropping any remnants of IE6 support. It goes on the forget pile.

[mathiasbynens]

+1 as long as we make it very clear that v3.0.2 can still be downloaded if IE6 support is a requirement.

[necolas]

The alternative is that we have a separate branch that isn't actively developed but which can be patched if necessary (or get jQuery updates if you really want).

[kristerkari]

+1, IE6 is dead everywhere else than China. What about IE7, is it going to die soon? In my country it has had only 2% share last 30 days: http://gs.statcounter.com/#browser_version-FI-daily-20120315-20120413-bar

[nexik]

+1, IE6 support should be threated as something special. In my country (Poland) IE6 0.65% and IE7 is on 1.91% of market share. Most people use IE8 (~10%) which is decent browser.

I know that in some countries IE7 is big market share, but IE6 should be threat as history.

[jonathantneal]

@necolas I think you and I have talked about how we (secretly?) maintain our own IE8+ versions of this CSS already. If we migrate away from IE6, can we jump to IE8? I think IE6 compatibility and usage is a bigger deal than it is for IE7 users since those users have already been interested in upgrades. Plus, then we can drop all of those asterisk hacks.

[tchalvak]

A branch seems like an excellent approach to otherwise dropping ie6, plus then patches (or lack thereof) can be considered votes for continuing to keep the branch around vs. Eventually dropping the branch altogether at some point.

[tchalvak]

And just curious, any reputable sources for ie7 browser share statistics? Wikipedia wasn't helpful, and w3fools have a pprobable userbase bias...

[alilleybrinker]

@tchalvak, StatCounter has a pretty great little interactive graph that has exactly what you're looking for: http://gs.statcounter.com/#browser_version-ww-monthly-201103-201203

Click on the browser names in the sidebar to show/hide the various browsers on the graph.

Based on the data provided, IE7 is currently at 2.91% market share worldwide.

[necolas]

@jonathantneal That thought had crossed my mind. But as Andrew points out, IE7 market share is still a little more significant. If we employ a methodical approach to phasing out explicit browser testing/support - say, Yahoo's < 2% market share criteria - then it's a bit early to phase out IE7 support too. But hopefully that isn't too far off either. The major leap forward will be when IE8 market share is below 2%.

[Inkdpixels]

+1 - Especially because Microsoft dropped their support for IE6 about a year(?) ago ("IE6 Countdown").

[tjvantoll]

+1

@Inkdpixels I believe IE6 will have extended support from Microsoft until April 2014, but still, it's time.

[mikefowler]

Agreed. I think a branch would be the easiest solution to provide the current code to those who still need it, but maintained support for IE6 seems an unnecessary burden at this point.

[mikealmond]

+1

[albohlabs]

+1

[roblarsen]

+1 20% of China is a pretty big number but if we let that number hold the project back we might never cut the cord.

[dmethvin]

What exactly is "dropping IE6 support" for H5BP?" Does anyone have a pull request with the changes? Or is this simply a "we won't test with IE6 and/or consider IE6-specific fixes from now on" thing? I ask because I am involved with a project that may be facing a similar decision soon.

[alilleybrinker]

I think it would include:

1. Removing the prompt to install chrome frame in index.html
2. Not allowing any more code that panders to IE6 bugs to be included in the project.

Sadly, much of the IE-related bug fixes in style.css are required for IE7 as well as IE6, and so they can't be removed just yet. The big moment will be dropping IE7 support. I count a total of 5 lines of code in style.css devoted to IE6&7 related problems.

[necolas]

We'll work out what to change later.

[joeybaker]

just to pile on… +1!

@necolas: where would be an appropriate place to discuss what to change?

[jonathantneal]

Compare the difference between normalize.css when IE6 is not supported and then when IE7 is not supported. In css normalization, the bigger "win" actually happens after IE7 is dropped. That's really interesting, because in practice you always think that IE6 is what will really bring you down.

necolas/normalize.css@ec330d5

necolas/normalize.css@e00d7a1

And @necolas, fear not, those are not real commits to master.

[robost]

+1

[yanneves]

I like the idea of applying print stylesheets to < IE7, but removing < IE7 support entirely works just as well. It's in good time.

+

[Gavrisimo]

+1

[necolas]

To be honest, we could consider dropping IE6 and IE7 if we're going to have a legacy branch.

[joeybaker]

I think it's a question of reasonable defaults: is it a good default to only support IE8+? If so, then a legacy branch makes sense. If not, then I don't see the need to bifurcate.

FWIW, I like the idea!

[paulirish]

IE6, IE7, and FF 3.6 are all dead to me.

IMO legacy branch is mostly frozen unless there are high-priority fixes that are needed to be made. no idea what qualifies as one of those in html5 boilerplate...

[necolas]

Exactly, the legacy branch would only be for potential critical fixes and to make it clear for some people that there is a legacy option if they need it. Basically, once I've decided how best to git-juggle a no-really-old-stuff version of Normalize.css and the existing version, it can work its way up to this project. Would be good to hear any ideas you have about that, Paul.

We could cut one last stable release of HTML5 BP, and then the next one would not be concerned with IE6/7 FF3.6 etc. and rely entirely on the expectation of chromeframe use.

[Mustapha-elouafi]

+1 it is the best idea IE6 / IE7 is a very big problem for a lot of web developper.

[zdzichu3000]

Hi,
I would suggest using in the index.html file such a piece of code:

<!--[if gt IE 8]><x-->
<link rel="stylesheet" href="css/normalize.css">
<link rel="stylesheet" href="css/main.css">
<!-- ><![endif]-->

IE8 and earlier display unstyled page,a nd if You switch jQuery to 2.0.1 site will not work properly with old browsers. Of course it is necessary to leave Google Chrome Frame prompt.

[alejandroiglesias]

@zdzichu3000 sounds like a good idea if you're not going to give basic-level support for older browsers (which is also a good idea) to be obvious for the user that the site is broken on his browser.

[imetallica]

There is a project from Chinese government to make Ubuntu the official operational system in China, which ships with Firefox by default. So I don't think IE6 will still be a issue in let's say one or two more years

[varemenos]

I'm pretty sure it's a Ubuntu-based distro, not Ubuntu itself. So it's not
sure what browser will ship as the default for that distro.
It could be Gnome's epiphany or chromium.

On Sun, Jun 16, 2013 at 3:33 PM, Iuri L. Machado
[email protected]:

There is a project from Chinese government to make Ubuntuhttp://www.ubuntu.com/the official operational system in China, which ships with Firefox by
default. So I don't think IE6 will still be a issue in let's say one or two
more years

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/1050#issuecomment-19511687
.

Adonis K. - Digital Artist, Web Designer & Developer
http://varemenos.com

[FlashJunior]

Global browser average (statcounter, netmarketshare, akamai, w3counter)
https://docs.google.com/spreadsheet/ccc?key=0AvDPn6IGEDe4dE5xMXo4MUJEYXFWYnBabUdyMDdwcFE#gid=0

[ghost]

jQuery 2.x will be used after that change, or you will keep with the 1.x?

[drublic]

@erick-ribeiro We will not update to v2. Please see #1353.