Deprecate IE7 in the Admin

[nacin]

markjaquith:

Everyone hates IE7. It’s insecure. Let’s make it go away. Also, dropping IE6 didn’t give us much beyond goodwill, because most of the hacks we needed for IE6, we also need for IE7. So we could actually clean up our CSS a bit if we dropped IE7.

[azaozz]

Big plus from me :) However going through the various browser usage stats here: ​http://en.wikipedia.org/wiki/Usage_share_of_web_browsers it seems IE7 is at average 7-8% of all browsers used. Perhaps still a bit too high, may be better to keep it around till 3.4.

[ocean90]

Related: #17232

[jane]

We agreed a bit ago that this was something we'd think about for 3.4, not 3.3. Punting for that reason.

[azaozz]

Replying to ocean90:

Agreed with "3.4-early" depending on the global usage numbers, don't think it would need a patch or testing though :)

[ocean90]

2011: 7.00%
​http://en.wikipedia.org/wiki/Internet_Explorer#Desktop_Market_share_by_year_and_version

Oct 2012: 2.32%
​http://marketshare.hitslink.com/browser-market-share.aspx?qprid=2&qpcustomd=0

[ocean90]

​IRC chat from today

nacin: I can pull some stats from browsehappy. Will take a week or two to collect.

The plan would be to leave any IE hacks (ie.css) for now but just no further support.

[aaroncampbell]

Here are the stats that I was able to find:

• ​W3Counter shows 5.79%
• ​Click says about 2.5-3%
• ​NetMarketshare says 2.13%
• ​StatCounter says .72%

[nacin]

Based on a sample of data sent to ​http://api.wordpress.org/core/browse-happy/1.0/, collected over the last few days —

• IE is 14.1%
• IE7 is 2.0%
• IE8 is 4.7%
• IE9 is 6.1%

My thoughts: We haven't done a whole lot to make IE7 work in recent versions. In 3.5, the stuff we did for IE8 already got us 90% of the way there. In 3.4, we had the customizer require postMessage, and thus didn't support it for IE7. In 3.6, more work will be directly in the admin (versus a separate experience, like the media library and customizer), but once you do IE8, we're usually pretty close to IE7.

Our level of support for IE7 has been pretty simple: just make sure it degrades in a somewhat graceful manner. If gradients are busted, that's fine. Lack of support for alpha transparency doesn't bother us. Even just declining to show a feature in IE7 is fine. But it doesn't take much to ensure that we have no terribly broken interfaces with boxes floating off the screen. How much we actually fix up in IE7 should be a feature-by-feature call, and probably only after we've done the IE8 support.

We used to group IE7 with IE6, but it's also fair to group it with IE8. Both are fairly old browsers that have very similar quirks. It's why jQuery 2.0 is dropping (sans-shim) IE 6, 7, and 8, rather than just IE7 and older. Also, Windows XP is EOL come 2014, which kills the operating system base for IE8 and below, so we'll probably start to see enough shift by 2014 or 2015 to pretty much drop IE8 and below.

[helen]

​https://irclogs.wordpress.org/chanlog.php?channel=wordpress-dev&day=2013-02-20&sort=asc#m559725

[netweb]

Replying to aaroncampbell:

Here are the stats that I was able to find:

• ​W3Counter shows 5.79%
• ​Click says about 2.5-3%
• ​NetMarketshare says 2.13%
• ​StatCounter says .72%

Above Jan/Feb 2013, Below Oct/Nov 2013

• ​W3Counter shows 4.47%
• ​Click says about 1.5-2.5%
• ​NetMarketshare says 1.31%
• ​StatCounter says N/A

[chriscct7]

I thought IE7 was deprecated in 3.3 or 3.4 but can't find any proof of that at the moment.

[tyxla]

Replying to chriscct7:

I thought IE7 was deprecated in 3.3 or 3.4 but can't find any proof of that at the moment.

​https://codex.wordpress.org/Version_3.2 says:

Start End-of-life (EOL) cycle for Internet Explorer 7

[aaroncampbell]

I think we're done with this as a ticket, so I'm going to close it. At this point our basic stance is: we don't officially support IE7 in the admin, so we don't sink a lot of time into it.

At this point, the consideration has really moved past IE7 anyway, so here are some stats as they stand:
​W3 Counter September 2015 Market Share says both IE8 and IE7 are too small to list (definitely under 2% likely much smaller than that)
​Clicky has IE8 at 1-1.6% and IE7 at 0
​Net Market Share has IE8 at 11.7% (quite the anomaly it seems), and IE7 at .4%
​Stat Counter has IE8 at 2.2% and IE7 has dropped off

Considering both the existing percentages and that ​Microsoft will no longer be supporting IE 7, 8, or 10 as of January 2016, I think it's safe to say that we can focus on supporting IE9 & 11, but we'll always support older browsers when it's easy to do so (and not damaging to the experience of the majority).

[SergeyBiryukov]

From a ​recent discussion, for reference:

desrosj: Is there a list of browsers/versions that patches should be tested against somewhere?
...
helen: latest for firefox, safari, and chrome. i would go iOS safari in the latest 2 versions of iOS. IE 7+; however, 7 is basically “don’t crash the browser” and 8 is “functional but doesn’t have to look good”.