I work on a B2B app with a lot of local govt users. We dropped support officially for IE11 last summer and I was surprised by the little to no pushback.
Our stance was if you needed to use IE11 for a legacy application, that's fine -- but our application required something besides IE and you're not limited to only one browser on your computer.
Any stances by IT that they haven't had a chance to authorize another browser or prove that another browser was secure compared to IE11 is an absolute joke at this point. Any IT department that is telling users to use IE11 for security reasons is questionable imho.
They learn from the best: a certain “open-source project to help move the web forward” has a bug tracker that simply doesn't load anything without Javascript, or in incompatible browsers.
It is funny how “moving forward” without brakes tacitly demands that others must clear the path and lay the tracks in order to prevent crashes.
Yep -- I thank those who fought that battle back in 2017-2019, as by time we got around to it, it was nothing. I bet we spent more time talking about if we should do it, when, and messaging than our support team has dealt with customers trying to use IE11.
It still comes up, we had a question this week from an IE11 user, and they just let them know to use another browser and they always do.
Hopefully with Microsoft's aggressive updating policies in Win10 old Edge will be killed off soon. We'll be sitting pretty when it comes to web standards. Safari lags behind but not by much.
In our team, every time there was a bug in Safari, it was something awfully wrong in our code that just happened to work on Chrome/FF for mysterious reasons. I never had a problem on Safari with standard quality code.
This is being far too kind to Safari. They were the ones who unilaterally decided that 7 days is sufficient time to clear out localStorage and IndexedDB.
While the File System Access API is still being developed, I'm not holding my breath for it to appear on iOS. To be fair, it isn't supported yet on mobile Chrome (for Android obviously), but I expect it will be added quickly after the v1 of the API is finalized. I expect it will either never appear on iOS or it will take 5+ years from now...
Why is it that iOS Safari was the first to support date inputs back in 2007, but desktop Safari, using the same engine, still doesn’t support it some 14 years later?
Apple should ditch WebKit and adopt Gecko. Mozilla could use the funding. and the two organizations share similar philosophies on user privacy. It would also deal a significant blow to the growing Blink monoculture.
According to the book “Creative Selection” (by one of the original Safari devs). They tried with Gecko first, but the POC didn’t went far: the build system at that time was messy and they couldn’t get it work. So they switched to KHTML, because of the nice code base. That internal fork evolved to WebKit.
I remember peeking around the Gecko codebase when Firefox got popular - and yeah, it was pretty gnarly. I remember seeing related .cpp files in the same directory using different naming conventions, for example.
WebKit is close enough to Blink that this move would be bad for the ecosystem as a whole. More variety in web technology implementations makes it harder for any one approach to dictate the standards going forwards.
Yup, it’s a fork of KHTML, much like Blink is a fork of Webkit.
One thing that WebKit does better than other browsers on OSX is hooking into the native rendering APIs. That’s something that Firefox does a shoddy job at.
Heh, ya, same deal where I work. Dropped support I think two years ago now. Anyone who asks doesn't complain when they are told to use a different browser.
Gitlab may not be the best example of potential pushback because the userbase is largely technologists. With this particular group, deprecating IE11 support is "preaching to the choir".
If the $1,000,000 specialist machine you absolutely rely on needs XP or 9x, which is very often the case in medical or industrial scenarios, that's not an option. What you _can_ do is use an alternative to the built-in (un-updateable) browser: although you won't be able to get the most up-to-date version of any alternative browser, it's still way better.
If that's the case, I would argue you should only use that specialized machine for specialized needs and have a more modern system for regular internet usage. Firewall the specialized hardware from any non-whitelisted sites / network addresses with extreme prejudice. Why risk a million dollar system to open a funny email with cat pictures, oh no its a virus.
Your highly important specialist machine is probably remotely exploitable for as much damage as replacement cost if it also needs a web browser and an internet connection.
Also a browser that is years behind current has a trail of breadcrumbs in terms of fixed bugs for newer versions. It is probably nearly as bad as the built-in option.
> Your highly important specialist machine is probably remotely exploitable for as much damage as replacement cost if it also needs a web browser and an internet connection.
Correct. Those versions of Windows will not bet getting any crucial or critical updates.
You're better off trying to see if ReactOS / FreeDOS will run the damn thing.
The only reason for continued IE11 support is because it shipped with Window 10 and they guaranteed that core software would be updated for a lot of years.
Back in the day (2009ish), we dropped support for IE6 for all but a few sections of our website that were used by employees of financial institutions. Same story - chrome was just gaining popularity and it was an easy sell. We had this giant banner for IE6 users to download Firefox or Chrome. Saved us so much time in development and testing.
Oh, not for us. I think we dropped support in like 2012 our 13 ish and our clients lost their shit. So many of them would cite the fact that they had special activex plugins installed into IE which made them more "secure".
Dropping 8 was a lot easier to do and we probably could ditch 11 today without a fuss. But man were our clients attached to 6, it was nuts.
Urg... Even in 2012 IE6 was ancient. If I was doing freelance and a client wanted IE6 support, I'm pretty sure I'd quadruple my rate and let them know it will take ten times as long to complete. Fuck IE6.
Yeah, that was bad browser for a developer who wanted to code only by standards. But in the other hand, back then when IE6 was released - it was the best.
I think the following two fact that played a big role in it:
1) Microsoft has another "official" browser
2) Chrome "eats the world", many people use it at home and many sites only officially support chrome. In many ways chrome has become the new IE. (Sites relying on non standardized Chrome specific quirks and being broken on other browsers, especially if they are not at least partially chromium based.)
I recently got actually burned by this. I was implementing the redesign for a friend’s website. I was working in Firefox, cause that’s the browser I want to support. The page seemed pixel perfect to me. Then my friend told me there was a subscription form he didn’t want at the bottom of the page. What was my suprise when I found out that there was an entire partial with an html error in it, that I missed in code, which was rendered in Chrome, but it was not rendered at all in Firefox.
I've been a Firefox user since version 2 and experienced this kind of thing early on.
Years ago I was making a GeoCities page and thought it would be a great idea to add a little picture/icon that followed the pointer around on the webpage (because that was a "cool feature" offered on the site builder at that time). I couldn't get it to work no matter which picture I tried, and I kept trying to add them many more times before eventually giving up on it.
Some time later I wanted to show off the site to a family member on their computer. I pulled up Internet Explorer (because they didn't have Firefox) and navigated to my webpage, where I was greeted by a swarm of icons following the pointer around and slowing the whole computer to a drag.
Lesson learned, test on multiple browsers because people will probably view your site in multiple browsers.
I'm more concerned about older mobile browsers. Almost everyone can install firefox/chrome on older PC's, but plenty are still running old android or non android phones as their only means of a 'computer'.
I've got a couple old android devices that won't even load most websites these days. Hell, the play store doesn't even work anymore. It's similar on old apple devices, but not quite as bad.
Android is Linux, even if it doesn't follow the philosophy of your favourite Linux distributions. Maybe RMS had a point when suggesting people call it GNU/Linux. Maybe Android/Linux?
Yep: GNU/Linux is very different from busybox/Linux is still different from Android/Linux, because the userland is (often) more important than the kernel; a user of Debian GNU/Linux has more in common with GNU/kFreeBSD than Android/Linux.
Yes, it's a very different experience running Android with Amazon's setup on a Fire Device than on with Google Play Services, even if it's binary compatible.
It's also a symptom of a wasteful greed driven system. By abandoning old hardware the industry at large can quietly force people into 'regular' upgrade cycles. That's two years for their ideal consumers who trade in their old phones and get the latest. At most it's like 6-8 years before the hardware is simply too old to possibly keep running. Many frugal people run a cycle or two behind, eg they're on iPhones (or model year equivalents) 8-11 currently depending on their cycle timeline. There's no way out of the cycle. I had a few coworkers get dropped by their cell providers over the last few years as the carriers dropped support for cell phones they bought in the 20xxs.
Essentially this planned obsolescence is manufactured consent in the population wide crowd funding of cellular technologies by for profit (mainly public) companies providing essential communication services.
I'll leave the moralizing of this to philosophers, but all these dynamics and their externalities undeniably bear further scrutiny.
yes , i also use a old phone which is probably 5 year old which only supports 1 of 10 websites. The chrome crashes all the times that is why i always use pc. The mobile phone market is growing quickly and we have to updates every month. Even window's needs to update montly. But still there are lot of people who wont buy new phone's every year but smartphone brands will never understand.
How many people use 5+ or 7+ year old phones? I’m sure they exist somewhere but I don’t know any of them.
It isn’t hard to get a 2-3 year old phone for free. A lot of people buy a new phone every 2-3 years and many of those people will be happy to give away their old phone if someone they know needs one
It's still a great phone, runs everything I've run on it really well, and as far as I can tell it's working as well the day I bought it. Still receives firmware updates too.
Its data connection is still faster than any other 4G device I've used, and faster than my home internet connection. The battery still lasts all day, and the ludicrously high resolution OLED screen is still in perfect condition.
It's hard to see why I'd want to replace it before it breaks.
If it lasts, I'll probably keep it a few more years, and because of the good experience, if it breaks I might get the same 4-year old model again second hand.
Even if you gave me a new phone for free, I'd put it in a drawer because there's no obvious benefit and considerable hassle to moving over.
My previous phone was 8 years old by the time I upgraded, went straight from Android 2.1 to 7. My current one is 3 years old and I don't see any reason to upgrade anytime soon.
I do. Because I live in an area where I need to use CDMA. Sure, there is a couple of spots in town that have other cell phone networks as well, but in a lot of spots and basically all buildings you need to use Verizon (or one of the resellers) or be without a cell signal. And sure they have LTE as well, but AFAIK the initial connection to the network is always over CDMA. And except for google maps (that freezes my phone for minutes, but is easily replaced by Osmand) and the banking app for one of my banks (I'd probably use that once a year to deposit a check because THAT technology is also not dead yet), there is nothing that doesn't work well on my Android 6 Samsung J1 Luna.
Even ignoring non-Microsoft browsers, there's Edge as a still-Microsoft, still-Windows alternative to IE that can be remotely/centrally managed with the same mechanisms. There really shouldn't be any reason to stick with IE11 anymore.
The only real remaining argument I can think of for IE11 is if you still have business-critical, legacy ActiveX applications.
There is even IE Mode in Edge for enterprises that truly need something in IE. Automotive industry has a lot of these still. 10-20 apps that are interacted with only work with IE.
> I work on a B2B app with a lot of local govt users. We dropped support officially for IE11 last summer and I was surprised by the little to no pushback.
I think the big part is that Microsoft wasn't just aggressive in pushing Edge but _especially_ emphasizing that IE11 was nowhere near as secure. Large bureaucratic organizations have a lot of inertia but they're getting a lot of pressure, especially at the federal level, to step up security and that overrides a lot of the familiar stalling tactics.
When I first started with my current company, old IE support was driven by our customer’s IT departments not having other browsers installed.
We’re now sort of unofficially NOT supporting IE 11 in so much as new features don’t get tested until someone complains.
What’s been interesting is that all of our government employees’ IT do have Chrome, Edge Chromium, and sometimes even FF installed, but the individual users are still trying to use IE11, and often will not accept “you must use one of the other browsers that you have available” as an option.
We’ve even gotten support tickets from customers about “broken features” which then acknowledged in their email that “they tried in Chrome and it worked”. One incident even acknowledged they had to use Chrome for other internal apps, but preferred to use ours in IE.
We SHOULD be pulling our support fully for it, because literally every customer has their choice of browser installed.
But users are so used to their IE they refuse to use anything else, and my boss seems to be happy accommodating that.
I work at a large enterprise company for their front-end. We have noticed that US domestic IE11 usage has dropped significantly, but has remained constant in Asia. Unfortunately some of our larger customers require us to support IE11 until its official security EOL (2025).
in the past year, major sites in all categories from linkedin to twitter to skillshare to microsoft to adobe have all dropped IE, with Wordpress, Drupal and Vuejs making plans.
And has stayed relatively up to date. That was the biggest PITA with IE is that it would be years behind other browsers in what APIs it supported and you needed a big painful update to the next major version to get new stuff.
IMO, MS has really mismanaged the IE product. The fact that there is now a legacy edge is really silly, why did they fork that instead of just making one of the version upgrades be a "Ok, this is a new change to the rendering engine which is fairly major".
Their aversion to making anyone unhappy is just silly. Chrome and Firefox have done a much better job with major rendering/js engine upgrades saying "Sorry some things are busted now, get use to it".
My understanding is that they didn't have an option. IE has so many idiosyncrasies that they couldn't make it compatible with modern Web standards and simultaneously retain compatibility with sites that were built for IE.
Better to write a new browser so that enterprise clients can use both - one for their legacy apps and the other for new stuff and the modern web - while they get everything migrated over one at a time. Versus, changing IE in place would have forced clients to rewrite everything overnight, which would have been a much, much worse situation.
If anything, I think supporting FireFox is actually quite useful since can highlight areas where one is relying on some Chrome specific behavior. Usually, if something breaks on FireFox, FireFox is actually the one that is correct in my experience.
I think this is a key point. I don't personally care much about deliberately dropping IE11 support. But I do want to use fetch() and similar functions without shipping a polyfill. I want to get the code size reduction that comes with just setting ES6 as a compilation target. I don't have to think much about Firefox when doing so, but I do have to think about dropping IE11.
i'm aware. i was holding off until all of IE is unequivocally 1.9% before jumping the gun. but someone already went and pestered them today. there needs to be no debate.
I'm not even joking when I say this, for some companies, it would literally be cheaper to buy your IE users a new computer. Obviously, for enterprise customers, who are likely to be the majority of IE 11 users, you can't necessarily do that (though think about if you can).
You could have some kind of wrapper that installs legacy web sites as a separate app and starts a locked down embedded IE with a set of security policies.
Users are dumb thats why. We literally have to make 4-5 shortcuts with iexpore.exe website.net and another one with chrome.exe website2.net and rename them accordingly. Some vendors have their web app work only with firefox and an old version like 20+ /:.
Not to upset any dev but really trying to put some fancy stuff and a bunch of frameworks in a piece of software that does one thing and has to work 24/7 is really a bad ideea.
There's good reasons why these exist, and it's got zero to do with dev competence.
In 2005 your employer asks for internal app, says you've got to use kendo or some such UI framework. Or maybe they want an autocomplete drop-down and you picked one that in the future will turn out to be not standards compliant. Or you were asked to use the stylesheet from another project to get them to look the same and that's not compliant. But you don't even know, as it only has to work on IE8 as that's all that's available to your users. You make it. It works fine.
Fast forward to 2015, no-ones touched the app in 10 years, the SVN server the code was on has been thrown away, and it's either rewrite the whole thing or force everyone to keep on using IE8.
That's what generally happened, not dev incompetence.
>Users are dumb thats why. We literally have to make 4-5 shortcuts with iexpore.exe website.net and another one with chrome.exe website2.net and rename them accordingly.
Edge in Enterprise Mode solves this as you provide edge with a list of sites that need IE, then Edge will automatically open those sites in IE Tab (or in previous versions of Edge, redirect the user automatically to IE)
From what I have seen supporting one browser is easier and maybe more sane for the IT department. I've seen chrome standard but certain employees get IE 7,8,9,10,11 and they use it for everything.
No joke. If you think about the computers that would need replacement, say purchased in 2010, could literally be replaced with some microcomputer that fully wipes and resets itself between logins that runs some desktop linux. These things are capable of driving a screen, keyboard and mouse and can connect to the internet - we've boiled down what a computer is to a re-settable science at this point - there's no need for a big ass tower with fans in it anymore.
Its not the computer. WE have a application that only runs on IE11. Our computers are windows 10. We use edge/chrome for everything else but 1 application still runs on IE11.
I still use IE as my default browser (and Firefox as a fallback), for the following reasons:
– The font rendering works better for my eyesight.
– Killer feature: Ctrl+N/K clones the browser window/tab including the history. I.e. you can effectively fork the history. EDIT: And, equally important, opening a link in a new tab/window also preserves (clones) the history.
– Certain keyboard operations have better usability than on other browsers.
– The title bar + address/tab bar height height is smallest among all browsers (after some configuration).
I’m slowly getting used to the inevitability of migrating to Firefox. The font rendering and especially the clone-by-keyboard feature are the hardest part though. (And the fact that Firefox doesn’t allow yellow as a search highlight color.)
Fwiw Firefox's "duplicate tab" feature similarly clones history. It only exists for tabs though, not entire windows. Right-click the tab, it's in the context menu.
To be fair, on IE it doesn't clone the entire window, just the current tab into a new window. Still, that's my main use case, and AFAICT it's not possible in Firefox by keyboard or even with a single mouse action.
There's a mouse shortcut for Duplicate Tab in Firefox: middle-click the Reload button! It blew my mind when I read about it in some random thread a couple of years ago.
Thanks. Now that I looked into it again, the issue I have is not just the Duplicate Tab feature, but that opening a link in a new tab or window clears the history of that new tab/window instead of cloning the history of the source tab. that means that you can't quickly open a couple of links into new tabs/windows, with each of them preserving the current history. I don't think there's a fix for that.
Some UI features on Firefox can be customized using userChrome.css [1]. I bet you can make title bar as small as you like and change search highlight color. For font rendering you could create a bug report.
Regarding the search highlight color, Firefox has the (mis)feature that if the highlight background color is too close to the text background color (which is the case for yellow vs. white background), then it inverts the search highlight text and background color, i.e. you get yellow text on black background for the highlight, instead of black text on yellow background. The rationale is accessiblity, i.e. requiring a minimum contrast for the highlight background against the normal background. But for me that behavior is anti-accessibility, and it can't be turned off. The issue has been raised since the behavior was introduced in 2009, but the maintainers don't want to change it. The corlor inversion implementation also seems to be pretty deep in the stack, and tied to the HTML <mark> tag.
Regarding font rendering, there have been multiple issue reports over the years, although most users complain that Firefox doesn't render text like Chrome, which I tend to find even worse than Firefox. ;)
If you middle-click the refresh button in chrome it clones the current tab with history. You can also middle click the back button to clone the previous page in a new tab with history preserved.
Thanks. The main problem is that there doesn't seem to be a way to access it by keyboard. Secondly, it's also not possible to convert a tab to a separate window by keyboard. So what is a single keypress on IE becomes a cumbersome operation.
Then actually it might not help you too much... it's really a replacement font rendering engine for Windows (initially Windows GDI rendering and now they're adding in support for DirectWrite). It uses FreeType with various parameters to make the font rendering look more like macOS (in a nutshell: macOS doesn't use aggressive hinting to force the font shapes into the pixel grid, which makes it blurrier but more accurate on low-resolution screens; Windows does, and so it looks sharper on low-resolution screens).
However while looking around at a bug report to get MacType to work in Firefox [0] I came across a setting in Firefox's about:config [1] that can change the font rendering and that might help you out. Not sure about the history forking thing but there may be some Firefox extension to help with that.
Yes, but (a) there’s too much software I still use that isn’t hi-DPI capable, and (b) I use a 1920x1200 display so would lose 10% vertical height, which is quite significant.
I've heard that the people who use IE11 are the ones who don't have a choice, because maybe they are using a computer at a library, or some kind of accessibility-related device. Does anyone on HN know of solid data to back this claim? This screenreader[1] survey for example says 11% of their users use IE11, though that is from 2019.
Look, something I’m finally qualified to answer! At my previous job our customers were a representative sample of the Fortune 500 and about 40% of our users (their employees) were on IE11 just a few months ago. There were some very well known companies that used IE11 exclusively. It’s very quickly changing and this might be the year when it finally goes away, but it’s still alive and well in the corporate world.
The biggest issue with newer browsers is actually probably the feature cycle. Every couple months browsers introduce at least one horrible feature I need to update our policies to disable.
I generally do the normal Firefox release and just put up with it, but I can understand why larger organizations might be frustrated with keeping modern browsers secure. They're worried about info leaks while browser vendors are interested in figuring out how to hook websites up to USB devices and VR headsets, and figuring out how to ad new advertising experiences to the new tab page.
Yes, considering the surface area they are relatively secure. But the surface area is absolutely massive and I have no doubt in my mind that there are multiple full remote code execution zero-days in all major browsers.
Treating security as an absolute is rarely useful because very few things are completely secure. No browser will be completely secure for the foreseeable future.
Huh, my last company was very successful in the enterprise space, selling to tonnes of Fortune 500 companies and government organizations, and IE11 was <2% of users for us. We were preparing to deprecate support for it around the time I left.
IE11: yes. IE11 "integrated into anything you were using IE11 for": no. MS is working hard to make sure that it fulfills its contractual obligation with respect to IE11's EOL, but it's at the same time removing it from everywhere it was traditionally integrated so that it can die a proper death. It'll be officially supported for as long as Windows 10 is a thing, but they're going to make it as useless as possible outside of "just being an old browser that's incompatible with today's web".
I still have IE installed so that I can run ONE god-forsaken enterprise application known as Oracle EBS (Enterprise Business Suite). For everything else I use Chrome.
Oracle EBS uses NPAPI (AKA Java Plugins AKA Java Applets). Chrome dropped support for NPAPI years ago, and since then we've been using IE to whenever we need to use Oracle EBS.
I understand that the latest version of Oracle EBS might now support the cutting-edge technology of Java Web Start to continue to deliver their 1990's era grey-blah UI's with shitty layout. I don't even care to ask about upgrades.
Negotiating with the bean-counter types that have enabled Oracle to put it's slimy tentacles into every critical area of the business is a dreadful, thankless task. I expect it will keep running, at its current version, long after I leave or retire. At some point, I expect employees will need to spin up VM's to open IE to use the f-ing thing-- perhaps they can set it up in their "forever" dream OS, Windows NT with IE4?
That's actually probably the best approach: make some Citrix hosts running IE11 and let people use those when they actually need to touch the legacy software. Then you can upgrade everything else.
It sounds like you (or your enterprise) has chosen not to upgrade you applications. That's fine. But it's hardly fair to blame the vendor for your choice of not upgrading.
I agree it's not entirely fair, but yes, the bulk of the blame has to go to the corporate derps that keep it running and love it.
But even if we could use the "latest version" a quick google image search for "Oracle EBS screenshots" shows that UI-wise, not much has changed. Maybe they got rid of the insipid flash-light icon? I didn't want to get too stressed by looking closely.
Easy reason: why spend money (and reduce quarterly profit) if the software is still supported by the vendor for another two years? Let it be the problem of the CTO of two years from now.
In two years the C-level/VP responsible has long since taken the golden parachute and his successor is stuck with the bill.
Also, any sort of internal IT stuff is exclusively seen as a cost center by the MBA bean counters. Stuff like employees leaving due to outdated IT isn't something quantifiable in impact unless you have entire departments walking away at once.
I used to work in IT at a Fortune 100 manufacturing company. IE versions hung around for a long time because there was a lot of software that was built with only IE in mind and only received the bare minimum of support.
In some cases this was off the shelf software from vendors who had gone out of business. In other cases it was software built in house that would require a complete rewrite to modernize.
All of this software was tightly integrated into processes that were critical for the company to operate. While the cost of replacing the old software was a consideration, the biggest concern was the potential disruption to operations.
If a software change reduced productivity or even worse caused work to stop entirely, it meant millions in financial impact.
1) Sometimes companies go out of business or stop supporting\updating things. If you build critical processes around these outdated tools, it can be very difficult to change.
2) Users aren't savvy enough to know when to use Chrome, Firefox or IE. Some aren't even savvy enough to call things the correct names (looking at you Mozzarella Foxfire). So they just settle on on a lowest common denominator.
All Windows screen readers with any significant user base have supported Chromium and Firefox for a while now. The old EdgeHTML engine didn't work so well with third-party Windows screen readers, so some blind users may have held onto IE 11 because of that. But any Windows screen reader that works with Chrome also works with the Chromium-based Edge. And note that the WebAIM survey you released was from before the Chromium-based Edge started being pushed to the broader Windows user base.
Bottom line: Web developers should not feel obligated to keep supporting IE 11 for the sake of screen reader users.
We build websites for organisations in the UK who do a lot of work with the NHS, and we regularly still see ~10% visits from IE11 in Google Analytics. Organisations that don’t work with the NHS consistently show < 2%, so although I don’t have conclusive proof I’m pretty sure that’s what’s going on for us.
I wonder how much of a sampling bias there is. It's plausible to me that users of non-IE11 browsers are more likely to be using ad/tracking blockers which block Google Analytics. That could skew the relative population of "users who ran Google Analytics" towards more IE11 users.
Were you able to use other sources than Google Analytics to look at the browser population? I imagine working with the NHS that things like HTTP access logs are harder to wrangle (if they exist at all)
For clarity, we don't build websites for NHS bodies, we build sites for some external organisations that work with the NHS, so logs would be relatively easy if it felt worth it. GA is my only source, but my hunch is that use of blockers is pretty consistent across clients whose visitors are mostly from businesses or institutions, and I see this differential across those audiences. Having spoken to people who work in the NHS Digital team there are definitely pockets of IE use, although it's been getting better - the shift sped up quite a bit after the Wannacry attack.
We have basic first party analytics of the NHS users of our software (mainly just browser versions actually), and it's probably a little over half of them are using IE11 still. My impression from talking to some of them is that most of them now have both Chrome and IE11 available on their machines, and it's just which one they're using. I'm sure that's uniform everywhere though.
People using IE 11 “because they have to” is a hard thing to quantify without some sort of survey.
I say this because even though your anecdote rings true, I’ve seen many people in offices use IE 11 purely out of ignorance and comfort reasons rather than because they were forced. They had Edge, Firefox installed but stuck with IE because no one told them to stop using it.
We may be underestimating the portion of “if it ain’t broke why fix it” crowd are using IE 11 but it’s hard to say with statistical data alone.
Anyway, regardless of how it’s sliced, it’s low enough now for libraries to start dropping.
We're writing apps for mostly local government users, almost all of who use IE11 because it continues to have Java Applet support.
Some of these Applet-based things have been migrated to modern web, but many still haven't, since it is almost 15+ years of software development now needing to be re-written.
There have been discussions about Edge's "IE Mode" but unfortunately that only applies to domain joined machines, and they have a BYOD policy.
Can you not have the employees join the domain, like how some companies let you enroll in Exchange on a personal phone in exchange for allowing device administration/remote wipe?
As with a phone, if it bothers the employees (it'd certainly bother me) have them buy a laptop for work only.
I think you mean, have the employer buy that laptop/phone for them. The employer provides an, imo, unreasonable ask, then they should have to provide the alternative.
Biggest reason is corporate/government IT that's been slow to upgrade, possibly because they have intranet sites that don't support newer browsers. Accessibility is available (and better) in modern browsers.
I do some server work for a small company that sells a downloadable product whose target audience is ordinary people, probably not too technologically sophisticated when it comes to computers.
It is unlikely someone would purchase from anywhere than their home Windows PC, and I don't think many are using accessibility-related devices (the site was using a third party component that turned out to be terrible for accessibility, and it was years before anyone complained about it (and we then replaced it with one that was accessible)).
I just took a look at the logs for the past 12 months. Here are the relative numbers of successful orders by browser, normalized to Firefox on Windows 10 = 100.
On Windows 10: Chrome 602, IE11 116, Firefox 100, Edge 94.
On Windows 8: Chrome 44, Firefox 16, IE11 14, IE10 4.
On Windows 7: Chrome 76, Firefox 40, IE11 30, IE9 2.
Basically. I work in the B2B online education space and many large businesses use IE11 on hardware that supports other browsers just fine. Part of it is that they use decades-old intranets that only support IE officially so they were stuck. We noticed since around early pre-pandemic 2020 (and accelerated by the pandemic), that IE users are becoming Chrome and Chromium Edge users. Our pet theory is that the Microsoft EoL pushed IT departments to just make the jump and then WFH pushed many to new hardware where they might have a choice. Also our product is far worse in IE as Trident's JS Engine truly is horrific and our whole site is much faster in Chrome. We saw IE use drop to under 10% now from ~35% around two and a half years ago.
IE11 is over 10% on the sites I build. Even when IE11 becomes only .5% of my users, I will support it.
I work in healthcare, so I don't get to make lazydev excuses about time constraints and "edge cases." The people still using IE are the people who need my web sites the most.
> I work in healthcare, so I don't get to make lazydev excuses about time constraints and "edge cases."
If I were to read this statement uncharitably, I would conclude that there is plenty of money in healthcare to be wasted on supporting outdated technology, adding to the already outrageous costs of healthcare. Furthermore, if you're not lazy, you're probably creating unnecessary work.
> The people still using IE are the people who need my web sites the most.
From a holistic welfare perspective, giving these people a modern web browser would be better than maintaining their status quo.
I work in healthcare as well and the health IT landscape is horrifying. Hospitals not hiring software engineers is part of the reason you end up with thirty-person implementation teams from firms like Epic and GE trying to config their way out integration hell. Trying to integrate a new product via HL7 is a special hell I wouldn't wish on my worst enemies. It's the only standard I've worked with where nobody even attempts to follow the spec, and everyone else is just expected to conform to whatever arbitrary changes the hospital made (both intentionally, and accidentally). The vast majority of IT management is completely non-technical, especially - paradoxically - at larger institutions. At small community places the IT director is typically the person who was setting up the doctors' computers ten years prior. At huge practices it's someone with an MHA or who hasn't done anything remotely technical since the 90's.
I love working in healthcare but the hospital side is atrocious in my experience.
I know there is a CURRENT Microsoft product (for administration) that requires IE11. I do not remember what it is but I was really surprised when our Microsoft guru (he is really good) showed it to me and was not happy either.
That's the thing, the edge cases are a long tail of accessibility challenges, and most of today's developers are OK with just telling that 1% or 5% tail to pound sand.
Me personally, I want my sites to be accessible from libraries, that old device in the closet you're attached to, survival centers and shelters, old kiosks stuck in the corner of a community space somewhere, and all other kinds of scenarios I haven't even imagined, but I don't want to leave those users out in the cold.
My companies thin client has to support IE11 because we run on lots of lock down industrial control systems that run Windows 7 or Windows 10 IoT and cannot have any other browser installed.
I'm counting 0.9% under our users (which is of course a biased group, but at least near being representative for region, gender, age and education) since 2020-01-01, but only 0.2% since 2021-01-01. So it seems be dropping quite steadily.
Edit: a few of the IE11 user agent strings even mention Windows NT 6.1, and no x64 or WOW/Win64, so I guess they're still running 32 bit Windows 7.
I find it amusing that people complain about React not being stable enough, and yet it still supports IE9 as Angular and Vue both look to remove IE11 support.
We have found that refusing to support IE11 in healthcare environments has lost us 0 customers and given them a reason to move away from it. At this stage any company using it and the likes windows xp or windows 7 are just creating unnecessary security risks.
I think at this point it would be some US auto-body shop with IE6 hooked up to some internal software, or maybe one of those big Japanese companies that never changes anything at all.
More like Chrome is the Web Standard now. Generally stuff that runs in chrome, runs fine in Firefox and Edge, Safari can be a little weird sometimes but overall it's way easier these days.
There are some fine differences. Google likes to blur the line between actual Web standards (which have consensus via W3C or WHATWG), Chrome's prototypes that may inform future standards, and self-serving Google APIs that only have "Web" in their name.
Sometimes it's just Google releasing whatever they want, use it on google.com and youtube.com while serving slower/buggier fallbacks to others, so other vendors have no choice but to implement Google's non-standard invention to avoid looking broken.
Google gets away with this a lot, because there are also many actual standards that Safari is ignoring. Without following dozens of mailing lists and bug trackers it's hard to tell what Google is pushing for themselves (e.g. AMP was a motivation for many "standards" proposals), and where others are dragging their feet.
I think the point seen here is really that if the standards org doesn't impact the browsers and the browser in question has greater than 60% market share then that browser is the standards org and the W3C doesn't matter. Who is going to harm Google? How would the W3C enforce their standards? If most of your users are on Chrome and Chrome sets the feature set rather than the W3C, why wouldn't you just build to Chrome?
Browser vendors have always been driving the standards. W3C is correct to call their docs "recommendations".
There was a brief nice moment in history when WebKit, Gecko, Trident, and Presto each had enough market share that all the vendors had to cooperate.
Nowadays Trident and Presto are dead. Gecko is a great engine, but doesn't have enough market share to veto anything. So Blink can ship anything and claim it's supported "everywhere except Safari".
The W3C gave up on many web standards like HTML and their process is just rubber stamping whatever WHATWG decides are standards and WHATWG seems to just rubber stamp whatever Google wants (more often than not).
Yeah, somewhere in the late 2010s we seem to have lost interest in the idea of graceful degradation.
It's one thing for a website to look a bit ugly and actions to require a few more clicks when accessed with an outdated browser. It's a completely different thing when you're left staring at a blank screen. The latter is what happens when an entire frontend framework decides that they'd rather give you no experience than a degraded experience because they don't want to maintain polyfills anymore.
I do sympathize with the hate for IE, having suffered it for long enough. But once we've decided that IE users don't deserve anything, who's next? How about the hundreds of millions of third-world users with grossly outdated Android phones? How about disabled people in middle-income countries who can't afford to upgrade their screen reader? In the past, using a crappy browser was a matter of choice. Nowadays, the only people who use them are those who have no other choice.
First they came for IE, and I did not speak out ... you know how the poem ends.
I think you have a valid point. However, I'd like to offer a factual correction on this part:
> How about disabled people in middle-income countries who can't afford to upgrade their screen reader?
As far as I know, this was only ever an issue on Windows, and it's not anymore. IMO, the best third-party screen reader for Windows is NVDA [1], and it's free and open-source. Even a user stuck on Windows XP or Vista can get a version of NVDA from 2017 that works well with Chromium (assuming one can get a Chromium-based browser that runs on those old versions of Windows). On all other platforms, the screen reader is built into the platform itself, and updated along with it. Windows itself has had Narrator built in for a long time now, and Narrator in Windows 10 is getting good (disclosure: I was on the Narrator team at Microsoft for 3 years), but there are still valid reasons to use a third-party screen reader on Windows.
You are right, but we must remember that poor people in poor regions are disproportionately more likely to be stuck on very old hardware. According to StatCounter, 5.7% of PC users in Africa are still on XP or Vista, compared to only 0.6% in North America. That's almost 10 times as much.
Rallying against older browsers was once supposed to be a good fight against evil monopolists. Now it's about kicking away the ladder from people who are already the poorest and most powerless in the world, making them even less able to take advantage of the latest information and communication tools. And we're not even realizing what we're doing because we're so used to the first world fast upgrade cycle.
> How about the hundreds of millions of third-world users with grossly outdated Android phones? How about disabled people in middle-income countries who can't afford to upgrade their screen reader?
None of these are Internet Explorer, the problem in question. Chrome and its Play Store dependencies are backwards compatible to much older phones. Android browsers even years ago had better web platform support than IE11.
I fully sympathize and believe that web sites should be as compatible as possible. It's the web apps that fall apart once you take away modern functionality.
> In the past, using a crappy browser was a matter of choice. Nowadays, the only people who use them are those who have no other choice.
This really depends on how much their actual lives depend on being able to use the web. If they live in poverty and/or third world countries, are their daily lives impacted by the web? What part of their life depends on what site exactly? Sure, communication is relevant, but we're already saying they have smart phones and thus data connections, email, and whatever native chat apps.
I'm not saying we should leave them behind technologically or that they don't matter or shouldn't have access to more information or knowledge, I'm just simply being practical with what is likely the real impact to their life.
In the end, I think web developers need to stop using ridiculous frontend tech for what ultimately constitutes static text content. The web already excels at that. Servers cache. Browsers cache. We don't need Vue to render a blog or article. Please leave such tech for SPAs and "apps" rather than sites or pages.
> If they live in poverty and/or third world countries, are their daily lives impacted by the web?
It won't affect their daily lives directly, but being able to access the knowledge scattered around the web is very important to human development in general.
What I wrote above is a kind of slippery slope argument, but it's a slope that I'm sure we're sliding down very quickly. You can't even read a technical blog these days without loading a whole SPA framework. If we keep going down this path, it will legitimately begin to impact the ability of people with older devices to access any information at all apart from those delivered through a handful of walled garden apps.
> In the end, I think web developers need to stop using ridiculous frontend tech for what ultimately constitutes static text content.
That was when we had no choice but to force in hacks to make things work on browsers that didn't quite follow the same standards. You ended up only being able to use the lowest common denominator feature set or using many hacks & polyfills.
Better is to stick with standards and make the browser makers implement them properly. Of course things aren't perfect still (Safari has a reputation for being slow on the uptake so you can't use the latest & greatest if you need to support those users, and it is best to avoid the bleeding edge anyway (at least sticking as far back as the oldest LTS release of common browsers)).
The operative word here is “pages”. Developers of statics sites should try to support everything. For full-featured web apps on a budget, the challenges of cross compatibility can be insurmountable.
That's 87.86% of what users are using (per Wikimedia's stats). To support the 1.76% of IE users, it will require a disproportionate amount of work that could negatively impact other development that benefits many more people.
That's the reality: Development time is a zero-sum game. If you do this you aren't doing this other thing, and in this case we're talking about under 2% that likely should be discouraged from using it.
PS - And this coming from someone whose uses are primarily on IE11 still (see my other post in the thread). Just because I personally benefit from IE11 staying around a few more years, doesn't mean the entire web should bend to that. IE11 must die, we just cannot turn on a needle, and when Edge dropped support for Java Applets/Flash it forced tons of organizations into IE11 for many more years than they would have wanted.
Code to standards and expect the browsers to implement those standards reasonably.
Avoid the bleeding edge to support those held back at LTS versions of common browsers, or if you use new tricks make sure they don't make things unusable on older-but-LTS browsers. Whether you consider IE11 an LTS browser or not is a matter for you to decide (yes, as it will be getting security updates for some time still, or no as it is seven years old, not getting any feature/support updates, and support for it is deprecated by even its own manufacture's apps).
Support accessibility by not using fancy things for the sake of it, and if you do something fancy make at least a little effort to have things degrade gracefully for those with accessibility issues. A lot of sites/apps skip this step, but shouldn't.
The problem is that IE chains your feet compared to every other single browser. CSS Vars? Nope! JS performance that is 1/50th that of Chrome on the same platform? Yep! Weird-ass rendering choices that make zero sense compared to the W3C standards? MS has you!
Basically, IE11 didn't change with the web so it's really out there now. You can develop for Firefox and all the Webkit/Blink offshoots just fine, but IE holds you up.
I've been in front-end dev for about 15 years and I don't remember that. I remember developing web applications to work in browsers that led to a significant number of our business' conversions, and practicing progressive enhancement in general.
Not sure what you mean. This is way less of a problem as it was even 5 years ago. Even a modern tool like Babel, which was pretty much mandatory in webdev build is becoming less important given the pace at which browsers standardise specs nowadays.
Yup, and I don't want to go back to resorting to hacks to piles of hacks to support them all. It's good that most modern browsers just work for the most part these days.
I am working on a few different apps right now and none of them are targeting IE. It's a huge time suck and just not worth doing at the expensive of improving other parts of the application.
I started using the web in 1994, and no, I don't ever remember a time when people designed web pages to work on any browser.
In the mid-90s, the internet was filled with pages that had "best viewed with Netscape Navigator" or "best viewed with Internet Explorer" icons. Twenty-five years later that mentality hasn't changed, and I don't think it ever will. It costs too much to target more than one or two browsers.
You can't have evolution without dropping ties to the past from time to time. That's about time to drop IE11 to take full advantage of modern Web features.
How much more does the web really need to evolve though? Maybe IE 11 already had all the features that the web platform really needs, and the upgrade treadmill is now being driven by nothing more than our insatiable demand for convenience as spoiled developers. Pity the users who can't keep up with the treadmill.
I agree that for many use cases, IE11 featureset is fine. We jump on the SPA ship too often. I think that static content and server-side rendered apps are often the best option that is for some reason considered out-of-fashion.
However, for those who see a Web browser as an application runtime, IE11 should be out of question. Vue is an application framework, and as such, it should benefit from all the improvements in the Web platform implemented in the modern Web browsers.
Heh, so we really ought to move from MSHTML... We're embedding a web browser control in our app for flexibility in adapting the GIS documentation system to varying needs depending on customer. Some customers document more details on electricity meters than others, some tie them to other billing systems on the backend than others. Etc...
But it's actually working out great and we can as a bonus update these parts of the desktop app without even shutting it down. Our app has a web API that the scripts interact with.
The obvious downside is that it's using IE 11. WebView2 exists though which uses Microsoft's Chromium fork, but the control is only now reaching production quality. Like mere months ago. https://docs.microsoft.com/en-us/microsoft-edge/webview2/
But it'll probably be what we eventually do especially as it supports Windows 7 which is quite something for emerging technology from Microsoft. We need this because surprisingly many customers still have Windows Server 2012 installs here and there. Sounds insane to run 10 year old servers for security reasons but the amazing thing is that Microsoft will support these to late 2023.
For now we're in our most advanced scripts using Bootstrap Vue + VueJS 2 + a few polyfills and it's amazed me how seamless it actually is in developing on Chromium 90 and everything just keeps working and looking almost identical on IE 11. Interaction, model bindings, templating, everything. I totally did NOT expect this. I expected... "Support". I mean... This is some achievement. There's been the occasional, minor stumbling block every three months or so but still...
WebView2 is basically CEF for the most part though. I would love for Trident or EdgeHTML to be released as open source honestly. I'm sure there are complicated licensing issues that would prevent this though. I built and helped maintain a relatively complicated application that interacts via the COM interfaces of MSHTML (and the crazy print template stuff) which basically has to be scrapped eventually.
Even though I've more or less zero IE11 users on Upstract, I still make sure that everything renders and basically works on it — I think this defensive, vanilla, bare-metal, no-gimmicks mindset creates better and simpler products — in spirit of what old-school Web tech was about.
(Unless you're building something like Figma of course).
If you think IE11 supports "vanilla, bare-metal" JavaScript or CSS you are going to have a bad time. In fact, if you are building a very simple website most of your code complexity will come from having to support legacy browsers and working around their quirks.
Modern Javascript and CSS both have features that allow for much cleaner and more readable code. Supporting IE11 seems like a really janky way to enforce code standards.
That's literally what most folks have been waiting for for years, I'm pretty sure most web devs are keenly aware of this. However, do note that there is technically also still an IE7 version floating around due to Windows Embedded Compact 2013 not having reached EOL yet (its EOL date is 10/10/2023).
This is obviously great overall, developing for IE is a pain. One weird anecdote though, I have an old laptop that has very little memory, running Windows 10. And I swear, IE11 works better than Chrome and Edge for websites that support it. Chromium seriously seems to crash my computer because of some memory issue. Any idea why this happens (regarding IE performing better)?
I only know this because I maintain a CSS framework that supports IE11, so I have to use IE quite a bit.
I used to have a similar issue on my laptop. IE was the only browser capable of viewing YouTube videos at 1080p without stutter. Firefox would stutter, as would Chrome and Chromium Edge.
I've since given the laptop a clear out, reinstalling Windows 10 and installing Kubuntu as well. The issue has now gone away. Firefox now copes fine from either OS.
I'm currently building a website that uses Vue 3 in parts, as well as other features that don't work (well) in IE11 such as CSS Grid.
Would you recommend only displaying an error screen for any users who try to view the site using Internet Explorer or another old browser? Or should I let those browsers try to display what they can, and maybe also a warning message?
My two cents is it depends on the nature of the lack of functionality, is this just about not displaying information or is this going to cause data corruption (ie, if validation checks or expected ajax calls are not being made), if it's just the display of things I would probably go with a warning message.
Keep in mind that the code has to compile for checks to work. Eg a ‘let’ in the same script block will break the isIE check too and it’s easy to miss that someone broke it. How often does the ie11 code get tested on ie11..
Best to keep the IE 11 warning code in a separate (nomodule?) script file
My B2B app is completely broken in IE. I just let it stay that way. No warnings. Tech support knows to tell customers to use Firefox, Chrome, or Safari if they run into issues. It's such a rare occurrence that it is not worth the effort to detect an ancient browser and render a warning.
Our stance was if you needed to use IE11 for a legacy application, that's fine -- but our application required something besides IE and you're not limited to only one browser on your computer.
Any stances by IT that they haven't had a chance to authorize another browser or prove that another browser was secure compared to IE11 is an absolute joke at this point. Any IT department that is telling users to use IE11 for security reasons is questionable imho.