Define guidelines for minimum IE support

[isabel_brison]

IE usage is decreasing steeply, but is still just enough that we have to support it. One concern is that there is a disproportionate amount of screen reader users that still depend on it, as confirmed by the latest WebAIM survey: ​https://webaim.org/projects/screenreadersurvey8/#browsers

With this in mind, it would be good to set guidelines for the level of support required for IE, so as to allow us to safely make use of newer technologies while guaranteeing accessibility for IE users.

• If we allow the experience to degrade gracefully, what is the essential functionality that must be preserved?
• What level of testing should be required to make sure everything works on IE?

[SergeyBiryukov]

Related: #46015

[azaozz]

Thanks @isabel_brison, great idea to track the progress on this. Changing the ticket type to "task" as it is not milestone related.

For starters, here's some current info about IE11 as of March 2020:

• Total usage: 1.63% (statcounter.com) - 2.39% (netmarketshare.com).
• Lifecycle: as IE11 is part of Windows 10, the EOL is the same for both. The latest release of Windows 10, version 1909, will be supported until May 11, 2021 for for Home, Pro, Pro Education, and Pro for Workstations, and May 10, 2022 for for Enterprise and Education editions (support.microsoft.com).

[netweb]

Related: #48743 Consider adding a notice do discourage IE11 usage

[twocs]

I don't like supporting IE11 but rumours of its impending demise may be exaggerated.

Internet Explorer 11 is the last major version of Internet Explorer. Internet Explorer 11 will continue receiving security updates and technical support for the lifecycle of the version of Windows on which it is installed. ​https://support.microsoft.com/en-au/help/17454/lifecycle-faq-internet-explorer-and-edge

Microsoft will continue to support at least one Windows 10 Semi-Annual Channel until the extended support date of October 14, 2025. ​https://support.microsoft.com/en-au/lifecycle/search/18165

My interpretation is that this means that Windows 10 could be supported beyond 2025, and IE11 by extension. Some long term enterprise versions of Windows 10 are supported until 2029.

[anonymized_14934761]

Comments use passive listeners and degrades the performance on sites with large amount of comments.

That's one of the reasons why the Google's Lighthouse scores are low on sites with WordPress installed.

The situation is solved only after running a special type of CDN that loads the content in a special way.

You should ditch the support for IE11 and improve the performance.

With the new features that new browsers, OSes and devices introduce for accessibility reasons, IE11 for sure is actually a worse option for them.

Btw I am not sure if the 3.39% shown for IE usage by W3 Counter is real as it may count the IE mode on Edge as IE11. Typically this mode is used by institutions to browse special sites or internal resources.

However with the discovered vulnerabilities and things, MIcrosoft is pushing updates and governments would buy new licenses and upgrade very fast.

Also Microsoft granted free upgrades extensions.

A lot of state websites were vulnerable and they would be fastly refactored so the IE mode won't be used as much on Edge.

That means that fake usage percentage would drop considerably.

Reported IE11 usage is not real, it is based on large scale due to the IE mode on Edge for administrative and government purposes.

Low-income and first-Internet access in the Third world now happens primarily with mobile phones, these phones primarily uses Chromium based browsers and Firefox, not IE11.

So the idea for discriminating someone with low income is not true in that sense. IE11 users are not Third world users.

I don't think the percentage is enough for you to support IE11.

The sites may still be browsable to some level, but you shouldn't anymore support IE11.

Even big corporations like Microsoft cut support on things like the Windows 7 despite the large amount of users and incentivizes them to move on with some tricks like the newly introduced "bugs".

Stop using passive listeners so that mobile users could benefit from an improved performance.

We need to "push" things in some sense as otherwise we would always be stuck in progress.

Apple ditched 32-bit support for applications despite many people using apps that would probably never be updated. That happened both on their mobile and desktop platforms.

There is some sort of "courage" needed to push things forward.

The real base of IE11 which is not properly represented in the statistics is far smaller and you shouldn't take it into account.

[burtrw]

To help maybe quantify the number of people that may be using IE11 beyond the general stats mentioned above, here are the percentages of browsers used by visitors to edublogs.org in the month of February 2020. Edublogs gets global traffic, with a large percentage of those in schools around the world (which are notorious for having outdated hardware and require the use of IE).

Top 10 Browsers

IE Versions

I also note that IE use is only slightly down from 2.13% of all visitors to Edublogs in February 2019 to 2.01% of visitors in February 2020. Before digging into this, I expected the drop in usage to be greater than it actually was.

It might be useful to see similar browser usage from other larger WordPress sites? To me, 2% is not really that insignificant, unfortunately. That being said, the idea of being able to safely make use of newer technologies while still not breaking or making sites on IE usable seems reasonable, if it is technically possible.

[Ipstenu]

A concern about this is not just the volume of users but who they are.

The people stuck on IE are often in government or financial sectors and are prohibited by law from upgrading things. The people who use IE 11 aren't just your parents who won't upgrade a laptop, they're people who actually have no agency over this change.

Keep in mind that any change we make that makes WP unusable for them will have larger consequences.

Yes, it sucks a LOT, but having worked in financial for 14+ years before I ran away, they're simply not capable of upgrading, and will stay using out of date and insecure products because they have no choice. Whatever we do, we have to keep them in mind.

[anonymized_14934761]

People who work for the governments around the world or financial institutions are required by law to upgrade after the extended support ended, unless these institutions pay an astronomical fee which makes it only valuable for super extreme cases like some ancient military facilities.

A large portion of the counted as IE11 users are actually Edge users that are using the "IE mode" on it.

Microsoft and Apple did some nasty tricks in order to force obsolescence. I think counting the performance enhancement that WordPress would gain by stopping using passive listeners would be significantly bigger value than keeping IE11 support.

Even Facebook doesn't support IE11 anymore and we both know what kind of users are using it.

[Ipstenu]

People who work for the governments around the world or financial institutions are required by law to upgrade after the extended support ended, unless these institutions pay an astronomical fee which makes it only valuable for super extreme cases like some ancient military facilities.

This is weirdly not true for all nations. Heck, not even true for all states in the US. Keep in mind, EOL is *2022* -- while that's two years away, it's also possible that there will be extensions like there were for Windows 8.

In some cases, they are required by law to use hardware or software written by one and only one check processing vendor on the planet, who refuses to update. I wish I was pulling that example out of thin air.

(Facebook not supporting IE doesn't really count here, since that's ONLY for social networking and unlikely to be open source product a company uses daily -- in fact, I'd bet most companies wold be totes okay that Facebook's broken)

My point remains this: We need to be very thoughtful and careful about what we break and how we handle workarounds. It would not be a bad idea to have an 'IE Compatibility Pack' type plugin to restore old functionality for these people in terrible situations. We just need to make sure we give them agency to move forward that isn't just "Lulz sucks! get a new browser!"

[anonymized_14934761]

I agree that things should be compatible, but they also needs to move forward.

In the EU, countries are forced by law to renew licenses for softwares that are no longer supported.

Yet keep in mind that in the US lives just 4.25% of the world population. While in EU it's 9.8%.

What I am worried is that a lot of these IE11 users are actually Edge users under "IE mode" that is now used for such institutions.

The Facebook example was in order to show that even a company, which has a large portion of users with outdated machines is pushing things forward and I think that is from a long time ago.

In every move there are pros and cons, the reasoning here is to weight each of these and compare them.

In India for example there are 1.34B people which counts for 17.7% of the world population, far greater than the US or EU and greater even if you combine them.

These people there, have their first internet connected device to be primarily low and even budget mobile devices, most likely a smartphone that runs a Chromium based browser or Firefox.

On the other hand they have significantly slower internet connection that comes throw the telecoms as mobile data.

For these people 2G and low speed 3G are something regular, not to count the latency and slowness due to the increased usage from so many people.

For these people a faster loading page that uses passive listeners for comments would be far better than the IE11 support.

It's a numbers game, I wish you just open a bit and compare what I am trying to say with an open mind as sometimes pushing things forward will make things better for the majority of people around the world.

Kind Regards

[azaozz]

Replying to burtrw:

To help maybe quantify the number of people that may be using IE11...

Thanks a lot for the stats! :)

It's very reassuring to have first-hand confirmation.

I also note that IE use is only slightly down from 2.13% of all visitors to Edublogs in February 2019 to 2.01% of visitors in February 2020.

Yeah, this also confirms the global stats I've seen. And yes, 2% of users is a huge number when it comes to WP.

[azaozz]

Replying to Ipstenu:

A concern about this is not just the volume of users but who they are.

Yes, completely agree. Usually the government, education and financial sectors are affected the most. Also, as @isabel_brison mentions in the ticket description, a lot of screen reader users are still stuck in IE11 because of outdated versions of their screen readers.

A large portion of the counted as IE11 users are actually Edge users that are using the "IE mode" on it.

Yes, this is quite possible too although (afaik) "Enterprise Mode" in Edge (that uses IE11) is usually turned on only for local networks/intranet.

In India for example there are 1.34B people which counts for 17.7% of the world population...
...their first internet connected device to be primarily low and even budget mobile devices
...they have significantly slower internet connection

Right. Don't see how supporting IE11 would "break" WP use there. True, some js files will be larger and not all possible optimizations would be implemented, but the block editor and wp-admin are still very much usable even on a slow connection and underpowered devices.

The alternative is to "break" WP for people that are stuck on using IE11 for a good reason, like screen reader users and some from the government, education and financial sectors.

It's a numbers game, I wish you just open a bit and compare what I am trying to say...

@slkfsdf8y34ljhsfsdfkuhfkl84hj I can assure you that WP is very much open to all opinions and also to different ways of personal expression (as long as they remain constructive).

I would really like to drop IE11 support tomorrow if it was possible, but that would mean making life much harder for certain groups of users, most importantly screen reader users (~11%) and people that have no choice in the matter.

[collinmanderson]

By the way, as @twocs mentioned, "rumours of its impending demise may be exaggerated."

IE11 is bundled with Windows 10 version 1909 ('19 Sept) which is supported for almost 3 years (2022). However, Windows 10 version 2004 ('20 Apr) is about to come out, and will also likely include IE11 and will also likely be supported for about 3 years (2023).

Microsoft has not publicly stated any intent to stop including IE11 in new versions of Windows (which come out every 6 months), so the IE11 EOL date keeps getting pushed into the future. Knowing Microsoft's obsession with backwards compatibility, IE11 might as well be supported "forever".

So Microsoft's official support policy for IE11 is "at least ~3 years", and it likely will continue to be "at least ~3 years" for a very long time.

I'm not saying either way whether WordPress should phase out IE11 support, just that we can't look to Microsoft EOL dates as a guide. (Or if WordPress does look to Microsoft EOL dates for IE11, then WordPress should continue supporting IE11 for the forseable future, as Microsoft has no public plan to stop including IE11 in new versions of Windows.)

[isabel_brison]

The people stuck on IE are often in government or financial sectors and are prohibited by law from upgrading things.

Too true. I used to work for a major Australian financial news website and, as of mid-last year, about 15% of desktop traffic came from IE11.

We need to be very thoughtful and careful about what we break and how we handle workarounds.

Yes! This is why I think having clear guidelines would be so useful.

It would not be a bad idea to have an 'IE Compatibility Pack' type plugin to restore old functionality for these people in terrible situations.

Oooh, thanks for the suggestion @Ipstenu ! Definitely worth looking into.

My reasons for raising this now are mostly Gutenberg-related, but I'm raising it here as ideally we should have a single strategy for the whole of core.

• CSS custom properties (not supported by IE) are seeing some use in Gutenberg: ​https://github.com/WordPress/gutenberg/pull/20047
• Not all Gutenberg features are manually tested on IE before being merged. We have some basic automated checks for JS, and there's a ticket open for adding some e2e tests: ​https://github.com/WordPress/gutenberg/issues/16509.

Clear guidelines around what absolutely can't be broken will help with development, and can also help us define what we should be testing for.

[desrosj]

I just wanted to drop some historical references to connect the discussions from the last time browser support was adjusted:

• ​https://make.wordpress.org/core/2017/02/28/the-new-editor-and-browser-support/
• ​https://make.wordpress.org/core/2017/03/13/continued-discussion-on-browser-support/
• ​https://make.wordpress.org/core/2017/04/23/target-browser-coverage/

[sabernhardt]

As noted earlier, people still using Internet Explorer likely are not doing so by choice. Let's continue supporting the people for a short time.

Level of support

1. If a task is more difficult—or even impossible—for someone to complete simply because of the browser's limitations, that needs correcting.
2. If the layout simply does not show as intended, people can still report and/or fix those issues (perhaps at a lower priority than other bugs).
3. For preferences, such as using a certain color scheme, a good fallback should suffice.

I know the editor has periodic IE 11 testing at least once with every major WordPress version, and testing with each major plugin release could catch more problems.

Discontinuing official support

In a Make post, announce plans that regular IE testing will stop as of a certain major WordPress release. We do not need to prevent access to certain features in IE 11; simply make it clear that things could be broken. ​Microsoft notified a full year ahead of dropping IE 11 support with 365 apps (not Windows 10 yet), though I like the round number of WordPress 6.0.

I suggest ending support in two stages:

1. Back end: the editor and any admin page that requires more than subscriber-level access
2. Front end: including bundled themes, theme-related functions, the admin toolbar, login and the profile page

If contributors still want to test their patches in IE 11 after that, they can continue that without the requirement.

[kraftbj]

Proposal make/core post: ​https://make.wordpress.org/core/2021/03/04/discussion-dropping-support-for-ie11/

[sabernhardt]

Guidelines for (dropping) support are here:
​https://wordpress.org/news/2021/05/dropping-support-for-internet-explorer-11/