Ten Years On: The Conspiracy to Kill IE6
May 1, 2019 2:51 PM Subscribe
The year was 2009, and Internet Explorer 6 remained a plague on the web. Already nearly a decade old, it was hopelessly outdated, riddled with bugs and security holes, and long since superseded. Yet with nearly 20% market share, web developers everywhere were still under pressure to support its infuriating quirks. The tide finally turned when, out of the blue, YouTube added a banner announcing it would soon be dropping support for the obsolete browser. Ten years on, it turns out that banner was never approved by management. Chris Zacharias reveals how a rogue team of frustrated developers launched the last major offensive of the Browser Wars, and got away with it.
The plan was very simple. We would put a small banner above the video player that would only show up for IE6 users... The code was designed to be as subtle as possible so that it would not catch the attention of anyone monitoring our checkins. Nobody except the web development team used IE6 with any real regularity, so we knew it was unlikely anyone would notice our banner appear in the staging environment.
This is the point at which I actually lol'd. Well-played, kids.
posted by Snarl Furillo at 3:17 PM on May 1, 2019 [27 favorites]
This is the point at which I actually lol'd. Well-played, kids.
posted by Snarl Furillo at 3:17 PM on May 1, 2019 [27 favorites]
If the YouTube team had ever deployed valid, standards-compliant HTML, maybe they could have got it to work cross-browser more easily. Only one Google property has ever served HTML that passed validation, I won't tell you which one, but I was the person who got it passing. Shortly thereafter I moved to another team and it was soon serving tag soup again.
posted by w0mbat at 3:31 PM on May 1, 2019 [30 favorites]
posted by w0mbat at 3:31 PM on May 1, 2019 [30 favorites]
We had a recent project where a client saw IE7 in their analytics and felt that they had to keep supporting it. We drilled down into their GA account, saw it was all coming from one IP, looked at the url requested, and realized that it was essentially an old computer in a waiting room with that page as IE7's default.
We told them we could add 5 figures for the scope change to support IE7, or they could spend $400 to buy a new computer for the waiting room. They made the right choice.
posted by fatbird at 3:32 PM on May 1, 2019 [75 favorites]
We told them we could add 5 figures for the scope change to support IE7, or they could spend $400 to buy a new computer for the waiting room. They made the right choice.
posted by fatbird at 3:32 PM on May 1, 2019 [75 favorites]
I have a story from 2014 about browser issues, though it isn't about IE6.
The non-profit I was working for had been asked by the government, as part of a contract, to provide a service via phone initially, and then also by online chat. We didn't think this was a big problem, as we had already implemented some level of using an online chat platform for certain services already.
So I worked with our IT provider for our online chat end of things, and we set up some test environments, and I did demo sessions with the government folks. And I did tons of testing with the folks from the government - we tried different browsers (IE, Chrome, Firefox, Safari, maybe one other) at different version levels. We tried them on desktops and laptops and mobile devices to make sure they would work. They sent me questions about security and data storage, which I sometimes had to check with our IT provider to help me answer. They asked for modifications to the sign on process, and then asked for changes back to the original configuration. And so forth, it went on and on, and then I got an email with this question:
Would it support Netscape Navigator?
Now, I hadn't thought about Netscape Navigator for more than a decade, possibly longer, at that point. And I actually researched what was up with Netscape Navigator in 2014 (which basically was that AOL tried to revive it in 2007, and that ended in 2008). I debated reaching out to the IT provider with this question, but I finally settled for a response to the effect of: "No. And I would suggest we wouldn't want (this chat solution) to work with something that old, because it would likely introduce security problems."
They were satisfied with that answer, thankfully. I still wonder what the meeting or process was that prompted it.
Anyways, that's my tale from a strange moment in the evolution of web browsing. I feel like Metafilter is somehow the right place to share it.
posted by nubs at 3:48 PM on May 1, 2019 [33 favorites]
The non-profit I was working for had been asked by the government, as part of a contract, to provide a service via phone initially, and then also by online chat. We didn't think this was a big problem, as we had already implemented some level of using an online chat platform for certain services already.
So I worked with our IT provider for our online chat end of things, and we set up some test environments, and I did demo sessions with the government folks. And I did tons of testing with the folks from the government - we tried different browsers (IE, Chrome, Firefox, Safari, maybe one other) at different version levels. We tried them on desktops and laptops and mobile devices to make sure they would work. They sent me questions about security and data storage, which I sometimes had to check with our IT provider to help me answer. They asked for modifications to the sign on process, and then asked for changes back to the original configuration. And so forth, it went on and on, and then I got an email with this question:
Would it support Netscape Navigator?
Now, I hadn't thought about Netscape Navigator for more than a decade, possibly longer, at that point. And I actually researched what was up with Netscape Navigator in 2014 (which basically was that AOL tried to revive it in 2007, and that ended in 2008). I debated reaching out to the IT provider with this question, but I finally settled for a response to the effect of: "No. And I would suggest we wouldn't want (this chat solution) to work with something that old, because it would likely introduce security problems."
They were satisfied with that answer, thankfully. I still wonder what the meeting or process was that prompted it.
Anyways, that's my tale from a strange moment in the evolution of web browsing. I feel like Metafilter is somehow the right place to share it.
posted by nubs at 3:48 PM on May 1, 2019 [33 favorites]
Man, so much terrible customized corporate software used old versions of IE. Like, databases written by people who didn’t work there anymore, and CAD PLM stuff. It’s like all the Y2K-noncompliant COBOL stuff got rewritten to depend on IE6.
posted by Huffy Puffy at 4:04 PM on May 1, 2019 [6 favorites]
posted by Huffy Puffy at 4:04 PM on May 1, 2019 [6 favorites]
"Move fast and break things" is usually antisocial advice, but in the case of "move fast and break IE6" I think we can make an exception.
posted by clawsoon at 4:10 PM on May 1, 2019 [10 favorites]
posted by clawsoon at 4:10 PM on May 1, 2019 [10 favorites]
My hate affair with IE dates all the way back to 1996 or so when I discovered that the then-current version of IE (2? 3?) didn't even support tables.
Circa 2010 I still had to support IE 6/7, but thankfully was allowed to compromise the page design to do so, and we weren't using much Javascript aside from a little basic jQuery.
posted by neckro23 at 4:26 PM on May 1, 2019 [1 favorite]
Circa 2010 I still had to support IE 6/7, but thankfully was allowed to compromise the page design to do so, and we weren't using much Javascript aside from a little basic jQuery.
posted by neckro23 at 4:26 PM on May 1, 2019 [1 favorite]
Most of the Canadian government runs on IE11.
posted by quaking fajita at 4:29 PM on May 1, 2019 [6 favorites]
posted by quaking fajita at 4:29 PM on May 1, 2019 [6 favorites]
An absolutely ghastly amount of corporate whereeverville, and even stuff like the web interfaces for Big Time Cisco Hardware only work in IE11(or have issues which lead to "please use this in IE"). This actually applies to a ton of network, management, and backend web consoles for Big Time Stuff.
It makes my head implode every time i'm introduced to more of it. It's now almost 2020, and i started yet another job where this is still the case, and we're on nearly state of the art stuff.
The companies public facing sites of course support chrome and such, but the back end web things that keep them doing business? IE.
posted by emptythought at 4:42 PM on May 1, 2019 [8 favorites]
It makes my head implode every time i'm introduced to more of it. It's now almost 2020, and i started yet another job where this is still the case, and we're on nearly state of the art stuff.
The companies public facing sites of course support chrome and such, but the back end web things that keep them doing business? IE.
posted by emptythought at 4:42 PM on May 1, 2019 [8 favorites]
My hate affair with IE dates all the way back to 1996 or so when I discovered that the then-current version of IE (2? 3?) didn't even support tables.
IE 3 definitely supported tables, and even had one of the first implementations of CSS; for the time, it was a fairly forward-looking browser. Wikipedia says both IE 1 and IE 2 supported tables but it's very likely that the support wasn't great.
posted by Aleyn at 4:58 PM on May 1, 2019 [5 favorites]
IE 3 definitely supported tables, and even had one of the first implementations of CSS; for the time, it was a fairly forward-looking browser. Wikipedia says both IE 1 and IE 2 supported tables but it's very likely that the support wasn't great.
posted by Aleyn at 4:58 PM on May 1, 2019 [5 favorites]
You may work with Chrome and Firefox all day, but please reinstall IE, and then install Flash and allow this ActiveX plugin to run so you can take our company's mandatory online security training.
posted by hyperbolic at 5:05 PM on May 1, 2019 [112 favorites]
posted by hyperbolic at 5:05 PM on May 1, 2019 [112 favorites]
That is a hell of a good story. As someone who used to work as an engineer at Google.. my hat's off to them. I felt I had this kind of authority on my projects, but then my projects had tiny niche audiences and I was also terrified to use that kind of power. The fact they had an institutional backdoor allowing this is nuts.
Only one Google property has ever served HTML that passed validation
You wrote this story as a condemnation of Google's practices, but it reads to me like an indictment of the value of HTML validation. And no, making your HTML standards-compliant absolutely wouldn't guarantee that an old bad browser like IE6 wouldn't barf in all sorts of bad ways.
posted by Nelson at 5:13 PM on May 1, 2019 [11 favorites]
OldTuber granted you the ability to completely bypass the new Google-oriented code enforcement policies, enabling anyone holding it to commit code directly to the YouTube codebase, with only the most glancing of code reviews from anyone.At some point Microsoft themselves launched a microsite which was "stop using IE6". Anyone remember where / when?
Only one Google property has ever served HTML that passed validation
You wrote this story as a condemnation of Google's practices, but it reads to me like an indictment of the value of HTML validation. And no, making your HTML standards-compliant absolutely wouldn't guarantee that an old bad browser like IE6 wouldn't barf in all sorts of bad ways.
posted by Nelson at 5:13 PM on May 1, 2019 [11 favorites]
At some point Microsoft themselves launched a microsite which was "stop using IE6". Anyone remember where / when?
Yep, it was IE6countdown.com, but it's no longer live. Although the internet archive remembers!
posted by jeremias at 5:24 PM on May 1, 2019 [4 favorites]
Yep, it was IE6countdown.com, but it's no longer live. Although the internet archive remembers!
posted by jeremias at 5:24 PM on May 1, 2019 [4 favorites]
just to jump on this train: our (US government) clients use IE11 plus a very restrictive VPN/local network situation that leads to just about every moderately fancy thing we try to pull on their websites look like total garbage. because the local issues are so localized (we have tried so many ways to replicate it but we can't because we're not literally inside their building) we basically have to rely on them to do our staging environment QC, which is kind of a nightmare because they, you know, aren't actually trained on how to do anything like that
posted by Kybard at 5:26 PM on May 1, 2019 [7 favorites]
posted by Kybard at 5:26 PM on May 1, 2019 [7 favorites]
Thanks! That launched in 2011 or a couple of years after this YouTube stunt.
posted by Nelson at 5:26 PM on May 1, 2019 [1 favorite]
posted by Nelson at 5:26 PM on May 1, 2019 [1 favorite]
> At some point Microsoft themselves launched a microsite which was "stop using IE6". Anyone remember where / when?
2011
posted by ardgedee at 5:34 PM on May 1, 2019 [2 favorites]
2011
posted by ardgedee at 5:34 PM on May 1, 2019 [2 favorites]
And nothing of value was lost.
posted by SansPoint at 5:35 PM on May 1, 2019 [3 favorites]
posted by SansPoint at 5:35 PM on May 1, 2019 [3 favorites]
Our timesheet software only runs on some version of IE with a Java plugin so putting my bi-weekly hours in requires me to fire up the Citrix receiver on my Mac and load a version of IE that our IT department had to setup just to be able to run this one application.
posted by octothorpe at 5:50 PM on May 1, 2019 [6 favorites]
posted by octothorpe at 5:50 PM on May 1, 2019 [6 favorites]
Metafilter discussing the YouTube banner.
posted by ardgedee at 5:59 PM on May 1, 2019 [16 favorites]
posted by ardgedee at 5:59 PM on May 1, 2019 [16 favorites]
We just had a very high up person tell us that IE11 is no longer supported for his Cabinet level agency. I think I saw people weeping.
posted by 922257033c4a0f3cecdbd819a46d626999d1af4a at 6:01 PM on May 1, 2019 [1 favorite]
posted by 922257033c4a0f3cecdbd819a46d626999d1af4a at 6:01 PM on May 1, 2019 [1 favorite]
What is dead may never die.
posted by sallybrown at 6:21 PM on May 1, 2019 [7 favorites]
posted by sallybrown at 6:21 PM on May 1, 2019 [7 favorites]
Only one Google property has ever served HTML that passed validation
You wrote this story as a condemnation of Google's practices, but it reads to me like an indictment of the value of HTML validation. And no, making your HTML standards-compliant absolutely wouldn't guarantee that an old bad browser like IE6 wouldn't barf in all sorts of bad ways.
Having also worked on the code of two different very popular browsers I can tell you that valid HTML is enormously more likely to render correctly on a randomly chosen browser than the crappy gumbo of data that most websites are.
The ways in which browsers cope with malformed HTML by doing things like guessing where to insert the missing close tags, guessing how to interpret things that are not escaped properly, how to handle text where the paragraph markup blocks are inbetween the paragraphs not around them, handle missing attributes that are not in fact optional - those are all things that are open to interpretation and vary between browsers.
That's not to say that browsers won't have different interpretation of valid HTML sometimes due to bugs or bad standard compliance, but it eliminates a whole giant source of randomness and dramatically reduces the headaches. It makes it so that you can at least file a bug report and say, "This correct, unambiguous HTML doesn't render according to the spec" and get it fixed.
posted by w0mbat at 6:33 PM on May 1, 2019 [13 favorites]
You wrote this story as a condemnation of Google's practices, but it reads to me like an indictment of the value of HTML validation. And no, making your HTML standards-compliant absolutely wouldn't guarantee that an old bad browser like IE6 wouldn't barf in all sorts of bad ways.
Having also worked on the code of two different very popular browsers I can tell you that valid HTML is enormously more likely to render correctly on a randomly chosen browser than the crappy gumbo of data that most websites are.
The ways in which browsers cope with malformed HTML by doing things like guessing where to insert the missing close tags, guessing how to interpret things that are not escaped properly, how to handle text where the paragraph markup blocks are inbetween the paragraphs not around them, handle missing attributes that are not in fact optional - those are all things that are open to interpretation and vary between browsers.
That's not to say that browsers won't have different interpretation of valid HTML sometimes due to bugs or bad standard compliance, but it eliminates a whole giant source of randomness and dramatically reduces the headaches. It makes it so that you can at least file a bug report and say, "This correct, unambiguous HTML doesn't render according to the spec" and get it fixed.
posted by w0mbat at 6:33 PM on May 1, 2019 [13 favorites]
Reading this I was going to express relief at not having to bear the burden of such a shifting landscape of compatibility concerns as befalls front-end developers, but then I realized that, in the last few months as a newly-minted (and thoroughly resentful) writer of Go, I've repeatedly had to deal with compatibility issues between minor releases of the core language and its associated awful tooling, and at least the JavaScript ecosystem is upfront with you about what you're getting into instead of gaslighting you into the notion that we've always been at war with go dep, and versionless packages, and whatever the fuck else.
posted by invitapriore at 6:49 PM on May 1, 2019 [3 favorites]
posted by invitapriore at 6:49 PM on May 1, 2019 [3 favorites]
and at least the JavaScript ecosystem is upfront with you about what you're getting into
Every time I roll up my sleeves and say, OK. let's learn some of this fancy Angular and React and Node and Vue and Bootstrap generation stuff, I start reading introductions and it seems pretty cool, and then I get to the bit where I have to run Babel and Grunt and other tools to make everything work in the browser, and I just die inside because it seems like it's going to be such a nightmare of versions and packages and mysterious errors.
posted by thelonius at 7:09 PM on May 1, 2019 [9 favorites]
Every time I roll up my sleeves and say, OK. let's learn some of this fancy Angular and React and Node and Vue and Bootstrap generation stuff, I start reading introductions and it seems pretty cool, and then I get to the bit where I have to run Babel and Grunt and other tools to make everything work in the browser, and I just die inside because it seems like it's going to be such a nightmare of versions and packages and mysterious errors.
posted by thelonius at 7:09 PM on May 1, 2019 [9 favorites]
Maybe those YouTube geniuses can turn their attention to their brilliant algorithm that thought I'd really dig a pro-David Duke video from someone called 'Be Open-Minded' presumably because I watched a clip of This is England yesterday.
posted by Alvy Ampersand at 7:40 PM on May 1, 2019 [20 favorites]
posted by Alvy Ampersand at 7:40 PM on May 1, 2019 [20 favorites]
You wrote this story as a condemnation of Google's practices, but it reads to me like an indictment of the value of HTML validation. And no, making your HTML standards-compliant absolutely wouldn't guarantee that an old bad browser like IE6 wouldn't barf in all sorts of bad ways.
Given that Edge, a good browser saddled with a logo that made power users reflexively shudder, recently gave up and used Google's rendering engine because Google kept publishing malformed HTML that only worked right in Chrome, and keeps publishing tools that only work right in Chrome, and then immediately afterwards broke ad-blocking in Chrome, I'm not sure 'well I guess standards don't matter' is the correct takeaway here. Standards are what stops the web from being shaped for the business interests of whoever owns the browser.
posted by Merus at 7:57 PM on May 1, 2019 [16 favorites]
Given that Edge, a good browser saddled with a logo that made power users reflexively shudder, recently gave up and used Google's rendering engine because Google kept publishing malformed HTML that only worked right in Chrome, and keeps publishing tools that only work right in Chrome, and then immediately afterwards broke ad-blocking in Chrome, I'm not sure 'well I guess standards don't matter' is the correct takeaway here. Standards are what stops the web from being shaped for the business interests of whoever owns the browser.
posted by Merus at 7:57 PM on May 1, 2019 [16 favorites]
Safari is today's bad browser. It's not as bad as IE6 was, but it's bad. And while having various options is good, assuming they all implement the standards correctly, being forced to use a bad option because walled ecosystem is pretty garbage behavior. It was with IE and it is with Safari.
posted by maxwelton at 9:30 PM on May 1, 2019 [2 favorites]
posted by maxwelton at 9:30 PM on May 1, 2019 [2 favorites]
i like safari but i haven't used chrome regularly in several years, it's such a memory hog and i no longer trust google enough to use it. there just isn't really a good alternative to those two on macs, as far as i know.
posted by JimBennett at 9:41 PM on May 1, 2019 [1 favorite]
posted by JimBennett at 9:41 PM on May 1, 2019 [1 favorite]
I feel like the "valid markup" discussion is kinda missing the point. HTML 5 does a reasonably good job of specifying how botched markup should be interpreted, and these days, if a reasonably standards-conforming browser is prevented from running a site when it works fine in another, it's usually either because of intentional (and some would say malicious) User Agent sniffing, or use of CSS/Javascript/DOM standards that haven't been widely implemented yet. I'm all for valid markup, but it's not going to solve all our interop woes, not by a long shot.
posted by Aleyn at 9:49 PM on May 1, 2019
posted by Aleyn at 9:49 PM on May 1, 2019
HEROES.
posted by DirtyOldTown at 9:53 PM on May 1, 2019 [1 favorite]
posted by DirtyOldTown at 9:53 PM on May 1, 2019 [1 favorite]
i like safari but i haven't used chrome regularly in several years, it's such a memory hog and i no longer trust google enough to use it. there just isn't really a good alternative to those two on macs, as far as i know.
Firefox, which continues to improve after the major changes a few years ago. It even has the advantage that extensions like uBlock Origin work better in Firefox than in Chrome due to API limitations, and the new Safari doesn’t even have the hooks to allow a new version of uBlock Origin at all.
posted by D.C. at 11:56 PM on May 1, 2019 [8 favorites]
Firefox, which continues to improve after the major changes a few years ago. It even has the advantage that extensions like uBlock Origin work better in Firefox than in Chrome due to API limitations, and the new Safari doesn’t even have the hooks to allow a new version of uBlock Origin at all.
posted by D.C. at 11:56 PM on May 1, 2019 [8 favorites]
That's not to say that browsers won't have different interpretation of valid HTML sometimes due to bugs or bad standard compliance
Web developers hated IE6 because it was nothing but bugs and bad standard compliance. You'd put together a simple enough page, test it in other browsers which would display it pretty much the same, and then fire up IE6 and see if barfed all up the wall.
We're not talking, "OMG, it's a few pixels wrong, my beautiful design is ruined!", we're talking the page being rendered unusable, with text sprinkled in random corners, layout rules interpreted in totally different, illogical ways.
It wasn't unusual to have to recode everything from scratch, using various IE6 cheatsheets that identified how to trigger which bugs to manipulate its layout engine into behaving predictably. While not changing how it renders in other browsers.
posted by grahamparks at 12:18 AM on May 2, 2019 [15 favorites]
Web developers hated IE6 because it was nothing but bugs and bad standard compliance. You'd put together a simple enough page, test it in other browsers which would display it pretty much the same, and then fire up IE6 and see if barfed all up the wall.
We're not talking, "OMG, it's a few pixels wrong, my beautiful design is ruined!", we're talking the page being rendered unusable, with text sprinkled in random corners, layout rules interpreted in totally different, illogical ways.
It wasn't unusual to have to recode everything from scratch, using various IE6 cheatsheets that identified how to trigger which bugs to manipulate its layout engine into behaving predictably. While not changing how it renders in other browsers.
posted by grahamparks at 12:18 AM on May 2, 2019 [15 favorites]
I remember once, after a full day of troubleshooting, figuring out that for the page to render in IE6, the title tag had to be above the meta description tag.
posted by Nothing at 3:13 AM on May 2, 2019 [6 favorites]
posted by Nothing at 3:13 AM on May 2, 2019 [6 favorites]
Being strict with standards only went so far, because IE6's renderer was built on a badly half-assed notion of how HTML and CSS ought to work. Floating anything in IE6 -- and at the time, if you wanted a multicolumn layout without using tables, css float was a mandatory tool -- would lead to nightmarish hours of increasingly hacky workarounds to cope with intrinsic IE6 rendering errors. And there were certainly many problems besides css float. And, of course, this doesn't even touch on the problems IE6 had with standard (at the time) Javascript.
By the late '00s, supporting IE6 typically doubled the time needed (at a minimum) for frontend development of anything that was expected to look and behave like a contemporary website; it was becoming an immense labor sink.
IE7 was also bad. To be sure, a much-needed improvement over IE6, but it also introduced a variety of brand-new problems as well as retaining some of IE6's worst idiosyncrasies, such as a box model for which a box's width value is the total of its contents plus padding plus border. All other browsers understood width to only apply to a box's contents by default, so HTML had to be written in such a way that static width values were only applied to block elements without borders and padding, or else IE7 and earlier had to be provided their own stylesheets. (A lot of people seem to think that IE7's behavior is how it's supposed to work, and to this day set box-sizing: border-box as a default behavior in their front-end frameworks. As far as I'm concerned it was never more than a hack to cope with IE6 and IE7, and is one reason why I don't like Bootstrap.)
posted by ardgedee at 3:36 AM on May 2, 2019 [5 favorites]
By the late '00s, supporting IE6 typically doubled the time needed (at a minimum) for frontend development of anything that was expected to look and behave like a contemporary website; it was becoming an immense labor sink.
IE7 was also bad. To be sure, a much-needed improvement over IE6, but it also introduced a variety of brand-new problems as well as retaining some of IE6's worst idiosyncrasies, such as a box model for which a box's width value is the total of its contents plus padding plus border. All other browsers understood width to only apply to a box's contents by default, so HTML had to be written in such a way that static width values were only applied to block elements without borders and padding, or else IE7 and earlier had to be provided their own stylesheets. (A lot of people seem to think that IE7's behavior is how it's supposed to work, and to this day set box-sizing: border-box as a default behavior in their front-end frameworks. As far as I'm concerned it was never more than a hack to cope with IE6 and IE7, and is one reason why I don't like Bootstrap.)
posted by ardgedee at 3:36 AM on May 2, 2019 [5 favorites]
This is glorious.
I get the feeling that there were a lot of us tactical-level web managers, at that time, all subtly trying to do stuff like this.
I was working at a major legal publisher at that time, for example, and IE6 support was the bane of our lives (law firms never upgrade their systems, or at least didn't then).
Then one day we were launching a new comparison product - probably the first in the sector to be properly mobile-first, responsive and... well... actually well designed and built around the user's needs (shocker). Early on though we realised we couldn't deliver half of what we needed to deliver in terms of USP, with the resource we had, if we had to try and make it work - in any way - in IE6. Yet we knew there was no chance even explaining that to the Senior Staff or the Sales Department, let alone the law firm reps they'd be selling to.
I told the team to not bother, but told them they had to keep a 'test version' that looked good and at least appeared to work with all the new features (in reality a lot, early on, were rigged demos) up at all times.
Then I told our MD that whenever he had a senior partner from one of the major law firms visiting, I would drop whatever I was doing and come demo 'where we were' with the new product for him/them. Which he loved because it made him him look techy and cool.
Each time, I'd demo it first on his machine (which I'd made damn sure got updated with the latest browsers), then offer to 'give access over our network' to the visiting partner (in reality, just browsing to the internal IP of our perma-test over the office wifi).
Almost every time it would then render like shite on their IE6-clugged laptop. But the thing was, they'd already seen it working fine, so then I'd just say:
"Oh, sorry, it looks like your IT guys haven't got around to updating your computer yet, maybe they've started from the bottom-up. That's a shame. Hopefully they'll have done it by the time you come back again, and then we can give you access for a proper sneak preview!"
After a while (and once it was robust enough for me to give the keys to the MD to play around with himself) I didn't even have to go to the meeting and give that line. The MD would do it for me.
The consequence: a shit-tonne of senior partners going back to their firms and demanding better browsers. Which said firm's IT departments then used to finally get the cash to actually upgrade stuff company-wide (at least that's what many of them told me later).
And by the time we finally launched that product, our IE6 user base had rapidly dropped to levels where I could legitimately claim the ROI wasn't there in supporting it.
posted by garius at 5:31 AM on May 2, 2019 [35 favorites]
I get the feeling that there were a lot of us tactical-level web managers, at that time, all subtly trying to do stuff like this.
I was working at a major legal publisher at that time, for example, and IE6 support was the bane of our lives (law firms never upgrade their systems, or at least didn't then).
Then one day we were launching a new comparison product - probably the first in the sector to be properly mobile-first, responsive and... well... actually well designed and built around the user's needs (shocker). Early on though we realised we couldn't deliver half of what we needed to deliver in terms of USP, with the resource we had, if we had to try and make it work - in any way - in IE6. Yet we knew there was no chance even explaining that to the Senior Staff or the Sales Department, let alone the law firm reps they'd be selling to.
I told the team to not bother, but told them they had to keep a 'test version' that looked good and at least appeared to work with all the new features (in reality a lot, early on, were rigged demos) up at all times.
Then I told our MD that whenever he had a senior partner from one of the major law firms visiting, I would drop whatever I was doing and come demo 'where we were' with the new product for him/them. Which he loved because it made him him look techy and cool.
Each time, I'd demo it first on his machine (which I'd made damn sure got updated with the latest browsers), then offer to 'give access over our network' to the visiting partner (in reality, just browsing to the internal IP of our perma-test over the office wifi).
Almost every time it would then render like shite on their IE6-clugged laptop. But the thing was, they'd already seen it working fine, so then I'd just say:
"Oh, sorry, it looks like your IT guys haven't got around to updating your computer yet, maybe they've started from the bottom-up. That's a shame. Hopefully they'll have done it by the time you come back again, and then we can give you access for a proper sneak preview!"
After a while (and once it was robust enough for me to give the keys to the MD to play around with himself) I didn't even have to go to the meeting and give that line. The MD would do it for me.
The consequence: a shit-tonne of senior partners going back to their firms and demanding better browsers. Which said firm's IT departments then used to finally get the cash to actually upgrade stuff company-wide (at least that's what many of them told me later).
And by the time we finally launched that product, our IE6 user base had rapidly dropped to levels where I could legitimately claim the ROI wasn't there in supporting it.
posted by garius at 5:31 AM on May 2, 2019 [35 favorites]
Our commercial bank (which will go unnamed to protect the guilty) requires IE for their commercial banking website. The online check scanning and such, it's all ActiveX stuff, and I can't even get it to work in Edge. I feel like one way or another, we're going to "need" IE for a long time. :(
posted by xedrik at 6:55 AM on May 2, 2019 [1 favorite]
posted by xedrik at 6:55 AM on May 2, 2019 [1 favorite]
Modern JavaScript is easy to make fun of (there's an npm module for that), but at least it's not ActiveX.
posted by tobascodagama at 7:27 AM on May 2, 2019 [1 favorite]
posted by tobascodagama at 7:27 AM on May 2, 2019 [1 favorite]
The online check scanning and such, it's all ActiveX stuff
RED FLAG RED FLAG
Oh my god I would not trust that institution with a penny of my money.
posted by arcolz at 7:43 AM on May 2, 2019 [4 favorites]
RED FLAG RED FLAG
Oh my god I would not trust that institution with a penny of my money.
posted by arcolz at 7:43 AM on May 2, 2019 [4 favorites]
so much terrible customized corporate software used old versions of IE
Two of my mission critical, daily-use applications not only require IE11, but require me to reset IE11 and "restore advanced settings" every single fucking time I log in.
posted by aspersioncast at 8:05 AM on May 2, 2019 [2 favorites]
Two of my mission critical, daily-use applications not only require IE11, but require me to reset IE11 and "restore advanced settings" every single fucking time I log in.
posted by aspersioncast at 8:05 AM on May 2, 2019 [2 favorites]
feel like one way or another, we're going to "need" IE for a long time.
Yep. I'm currently working in an organization that is health-care adjacent; whenever I'm invited to an online meeting with the folks in the health care system, they use Skype for Business and Microsoft Lync. It took me forever to realize that if I wanted access beyond the phone bridge (like, to see presentations or whatever), I needed to open the meeting space link in IE rather than FF or Chrome.
Someday, there will be a TV series (or whatever has replaced TV by then) about these early years of the internet and the kludge; I picture it being like Deadwood, where encroaching "civilization" somehow makes everything worse.
posted by nubs at 8:15 AM on May 2, 2019 [1 favorite]
Yep. I'm currently working in an organization that is health-care adjacent; whenever I'm invited to an online meeting with the folks in the health care system, they use Skype for Business and Microsoft Lync. It took me forever to realize that if I wanted access beyond the phone bridge (like, to see presentations or whatever), I needed to open the meeting space link in IE rather than FF or Chrome.
Someday, there will be a TV series (or whatever has replaced TV by then) about these early years of the internet and the kludge; I picture it being like Deadwood, where encroaching "civilization" somehow makes everything worse.
posted by nubs at 8:15 AM on May 2, 2019 [1 favorite]
there just isn't really a good alternative to those two on macs, as far as i know.
Firefox, which continues to improve after the major changes a few years ago.
This. Firefox is greedy about RAM on OSX, but around ten tabs it balances out compared to chrome, and stays fast however many tabs I open.
posted by aspersioncast at 8:30 AM on May 2, 2019
Firefox, which continues to improve after the major changes a few years ago.
This. Firefox is greedy about RAM on OSX, but around ten tabs it balances out compared to chrome, and stays fast however many tabs I open.
posted by aspersioncast at 8:30 AM on May 2, 2019
RED FLAG RED FLAG
Oh my god I would not trust that institution with a penny of my money.
I did financial IT for years. You better just put your money in your mattress then, because even if the banks customer facing site works in chrome, i guarantee you the day to day operations and internal processes require sites exactly like this. I literally had my keepass store divided between "REQUIRE IE/MUST USE IE" and other, with a small list of "broken in IE" too
The IE list was substantial, like, you had to use the scrollbar. I almost wept when i first saw my bosses, of course much more substantial, list.
posted by emptythought at 9:21 AM on May 2, 2019 [2 favorites]
Oh my god I would not trust that institution with a penny of my money.
I did financial IT for years. You better just put your money in your mattress then, because even if the banks customer facing site works in chrome, i guarantee you the day to day operations and internal processes require sites exactly like this. I literally had my keepass store divided between "REQUIRE IE/MUST USE IE" and other, with a small list of "broken in IE" too
The IE list was substantial, like, you had to use the scrollbar. I almost wept when i first saw my bosses, of course much more substantial, list.
posted by emptythought at 9:21 AM on May 2, 2019 [2 favorites]
If the YouTube team had ever deployed valid, standards-compliant HTML, maybe they could have got it to work cross-browser more easily.No. IE6 has never been a good citizen, and (as noted in the article) would fail spectacularly if it hit certain valid CSS or HTML directives.
Cross-browser compatibility is a great goal, but that does not and cannot include IE6.
posted by uberchet at 1:10 PM on May 2, 2019 [1 favorite]
I have suffered for so long. I was only allowed to stop supporting IE6 three years ago. This is on a sizable corporate website with two magazines, events, news, brochure pages, the works. Couple of thousand URLs all in.
When a sizable chunk of your customers are in Asia somehere and are still using Exploder, it's impossible to justify dropping it on the traffic numbers. In the end I built a new responsive home page with a slideshow and a fancy fly-out menu - the head honcho fell in love with the shiny when he saw it on his iPhone, and cheerfully threw IE6 under the bus.
It still makes me angry thinking about the development time I wasted supporting that horror.
posted by Gamecat at 1:45 PM on May 2, 2019
When a sizable chunk of your customers are in Asia somehere and are still using Exploder, it's impossible to justify dropping it on the traffic numbers. In the end I built a new responsive home page with a slideshow and a fancy fly-out menu - the head honcho fell in love with the shiny when he saw it on his iPhone, and cheerfully threw IE6 under the bus.
It still makes me angry thinking about the development time I wasted supporting that horror.
posted by Gamecat at 1:45 PM on May 2, 2019
« Older It was rather a long way in those days. | Behind the Cat Pictures Newer »
This thread has been archived and is closed to new comments
Our FAQ was simple and to the point:
Q) Why won't the training load in IE6?
A) Because you are using IE6.
posted by delfin at 3:08 PM on May 1, 2019 [41 favorites]