A Conspiracy to Kill IE6 (2019)

https://blog.chriszacharias.com/a-conspiracy-to-kill-ie6

by SerCe • 5 years ago

241 80 5 years ago

blog.chriszacharias.com

wolrah• 5 years ago

Proof that deprecating support for ancient software run by change-fearing businesses can still be done properly.

The people responsible for this are heroes, they did what even Microsoft themselves were unable to do without their help, and they made the entire internet undeniably better by doing so.

ksec• 5 years ago

What I a great story. This is from 2019 and it somehow fly under my radar. I was always under the impression killing IE6 was a strategic decision from Google especially it came after they introduce Chrome where they previously said they were 100% behind Firefox. And we have email evidence that is not the case as they had always intend to have their own browser.

It is also Interesting how Youtube seems to have resisted integration into Google infrastructure and tools for as long as possible. Vitess was created in 2010 to solve their mySQL scaling issues. And I believe it wasn't until late 2019 they moved it to some whatever Google standardise.

deaddodo• 5 years ago

YouTube has resisted it specifically because the spectacular failure of Google Videos.

But you can totally see Google's hands in their business decisions such as shoving GPM into YouTube Music and switching their entire ad stack to be Google exclusive (an obvious move).

tonto• 5 years ago

Corresponds with dropping ie11 today too https://twitter.com/cramforce/status/1443962459723755533?s=2...

fiddlerwoaroof• 5 years ago

I still remember the day this came out and how happy I was that I could now convince the project I was working on to give up on IE6 too.

tombert• 5 years ago

I wish I had thought of sneaking in a scary banner in ~2013 when I still had to support IE6 for a web development job I had.

I, like a majority of web developers, was using Chrome to do all my development and testing, and would only occasionally VM Windows XP to test Internet Explorer, which usually had something broken, and would usually end up with me spending a majority of my time fixing, and when I would complain about it, management would point to the server logs and say "Look, N% of people are still using IE6, we have to support it!" If I had been smart enough to put a little deprecation warning on there, life might have been easier for me.

irrational• 5 years ago

And who gave management the numbers from the server logs? You guys should have fudged the numbers.

solarengineer2• 5 years ago

I’ve replied elsewhere in how I reported the support overhead for older browser versions.

While one could fudge numbers once, we could get accustomed to misrepresenting in the future too. Once users/customers discover this, they’ll always keep second-guessing us. Trust once lost would be difficult to regain. It is better to how customers that they need not waste money anymore and shouldn’t support/facilitate users using insecure and outdated technology anymore either. They could give it a positive spin and get good PR too.

bpodgursky• 5 years ago

This is terrible advice. Taking a bold gamble without total authorization can pay off sometimes... lying to management is how you get fired (at best) or sued.

ffitch• 5 years ago

To be fair, IE6 share was falling quickly on it’s own, and its sharp one-week drop coincided with IE8 launch. IE7 also saw drop that week, even though it didn’t have the deprecation message. Google’s banner probably made IE6 die a few months sooner, but that’s it.

solarengineer2• 5 years ago

For years after, various customers facing websites required IE6 support because analytics showed that there was significant IE 6 traffic. Various enterprises were slow to move off IE6 because various websites and products continued to support it!

I’d learned my lesson. When IE 7 was similarly entrenched, I’d start to provide separate estimates for IE 7 support, and explain to client-side business and IT folk that they could use this information to justify developing reduced functionality for IE7 users and for showing such users a banner urging them to upgrade to a newer browser.

ahmedfromtunis• 5 years ago

That's a great story and a beautiful reminder that relying on data to make decisions is not always a good strategy.

I sometimes feel like people forget that data is an effect of our decisions and actions, not the other way around.

A bit off topic, but this is also a reminder that even conspiring to put a miniscule banner can't go unnoticed, let alone one with much grander objectives like inserting chips into billions of vaccines, or fake a moon landing.

noduerme3• 5 years ago

Excellent comment. Yes, it's terrifying to think of the social power of one silly banner on a huge website - that the bosses don't even notice! It's also really true, and I see it all the time on HN, that people expect data to make decisions rather than understanding that their own bias is shaping the data. Much of this ML data analysis approach is like shaping a huge statue on tiny legs, each person making it bigger and bigger at the top until it falls over.

The best thing about this story is it reminds you to be a rebel.

xg15• 5 years ago

A bit off topic, but this is also a reminder that even conspiring to put a miniscule banner can't go unnoticed, let alone one with much grander objectives like inserting chips into billions of vaccines, or fake a moon landing.

I'm sorry I have to defend the tinfoil hatters here, but I think it's important to note the alleged situations are different: In the OP, we have a small group inside a large organisation (Google) conspiring "against" the organisation (even if ultimately to its benefit).

Most conspiracy theories seem to assume the whole organisation (government, military-industrial complex, NSA/CIA/..., lizard people, etc) conspires against the general public. So, not so much an actual conspiracy, more like large-scale corporate projects which are hidden from the public and against the public interest. Inside the organisation, there would be no need to keep anything particularly secret as the conspiracy is a perfectly well-approved project of the organisation.

Of course stuff like fake moon landings and vaccine microchips are still bullshit - however, events like the Snowden leaks or numerous cold war CIA operations which have been confirmed by now show that in principle, it's possible to keep large-scale projects completely unknown to the general public for an extensive timeframe.

I think if anything, the OP could be a sign how hard it would be to reveal such projects. Because any potential whistleblowers would find themselves in a similar situation where they have to conspire against the organisation they are working in.

cletus2• 5 years ago

This is a good story. You can take away several things from it. Here's mine: it demonstrates the value of reputation and trust and how easily they're lost and how hard they are to recover.

IE4 came out at a time when Microsoft considered Netscape an existential threat. IE4 was really the last nail in the coffin. It was actually way better/faster than NS. By the time IE6 came around, there was no imminent threat to Windows in the form of the Web being a platform replacement. So IE6 stagnated and became the bane of the Internet. IE-specific hacks (and IE6 in particular) became the norm for Web developers. It was pervasive too. For example, IE had a completely different interpretation of the CSS box model to everyone else (specifically, how padding, margin and width should interoperate).

Fast forward some years and Microsoft is not the same place it was under Gates or, more importantly, Ballmer (he had the longer tenure in the modern Internet era). Nadella really does seem to represent a different approach and MS has produced some great tech (eg Typescript).

But IE died because of the reputation it earned and all the existing hacks around it. It got renamed to Edge specifically to avoid the if-IE blocks and to more cleanly cast off IE-legacy behaviour.

But the fact is, no one really trusts MS anymore when it comes to the Web (Typescript notwithstanding). So Edge just became rebranded Webkit much like Safari/Chrome.

To be fair, some other stuff contributed to the death of IE and some of it was pretty dodgy (eg little things Google did that crippled non-Chrome browsers such as hardware video decoding on Youtube).

Anyway, you see this too with Google and what's now a meme about chat applications (Meet, Talk, Hangouts, Duo, allo, anyone?). Google is also now a meme for dropping support for things people use. Reader is the posterchild for this even though my opinion is this wasn't as big of a deal as people make out.

But the cost in all that is now no one trusts anything Google launches to be there in 3 years. Reputation matters. Apple gets this even though they've also killed some things (eg Safari for Windows). Google really doesn't.

Lastly, as far as Youtubers having an identity of not being Googlers, this was certainly still the case 5 years ago and may still be the case today. They have a separate campus (in San Bruno) and generally eschew anything Google3-related. They also still called themselves Youtubers.

goldenkey• 5 years ago

I suppose it's kind of humorous then that YouTube has hardly been integrated into the Google ecosystem sofar as a redirect to auth.youtube.com is done on other Google products to propagate the auth information. This has caused issues with teachers wanting to block YouTube yet still have Google Classroom [1]

[1] https://news.ycombinator.com/item?id=28582334

cwyers• 5 years ago

IE4 came out at a time when Microsoft considered Netscape an existential threat. IE4 was really the last nail in the coffin. It was actually way better/faster than NS. By the time IE6 came around, there was no imminent threat to Windows in the form of the Web being a platform replacement. So IE6 stagnated and became the bane of the Internet. IE-specific hacks (and IE6 in particular) became the norm for Web developers. It was pervasive too. For example, IE had a completely different interpretation of the CSS box model to everyone else (specifically, how padding, margin and width should interoperate).

Well, let's also remember that the governments of the the United States and the European Union came in and intervened with Microsoft's use of Internet Explorer. With the benefit of hindsight, we can see this is an aberration -- every other major consumer OS (macOS, iOS, Android) has a bundled web browser that has significant market share among that OS's user base. No government seems to have gotten involved in these operating systems.

Microsoft's decision to stop investing so much in IE can be contextualized by noting that major world governments intervened to insist that Microsoft's marketshare in web browsers was reduced. It's not a shocker that organizations respond to their incentives, and Microsoft's incentives to spend a lot of money on IE were reduced by government involvement.

cletus2• 5 years ago

... every other major consumer OS (macOS, iOS, Android) has a bundled web browser

At the time Windows had a >90% market share and there weren't mobile Internet devices so it should be held to a different standard.

That being said, I do think governments rush to action on tech issues where it's completely unnecessary.

I do agree the agree the government suit did play a role in browser tech languishing at MS. I honestly believe they didn't think they were doing anything wrong. After all, this was part of MS's "embrace, extend, extinguish" MO. They tried the same thing with Java. So after that it felt to me (as an observer) that MS didn't really know what to do.

TedDoesntTalk3• 5 years ago

As I recall, in the US at least, Microsoft was not forced to unbundle IE from Windows. It was still the default browser and pre-installed.

Lio2• 5 years ago

From the outside it appeared very strongly that Microsoft deliberately allowed IE to stagnate with the end goal that the web should mean IE and IE only.

That wasn’t a response to government but a strategy to tie the web to Windows as the only platform which could run real IE6.

They could bundle a browser with macOS or Linux but it was much use if it wouldn’t work with your bank, corporate or government sites.

They used the same strategy to have Office files be memory dumps of the programs that created them and formidably hard to support by 3rd parties.

Macha• 5 years ago

For example, IE had a completely different interpretation of the CSS box model to everyone else (specifically, how padding, margin and width should interoperate).

One interesting outcome of this is we now have the `box-sizing` CSS rule to choose between the original, default box model (content-box) and the IE box model (border-box). Basically every project I've looked into opts into the IE box model.

sangnoir2• 5 years ago

The IE box model was correct, but different from all other browsers, and in addition to all the other bugs and IE6-isms (which were wrong), was pretty annoying to deal with.

i_like_apis• 5 years ago

Back when many people attempted to support ie6 and even ie4 (oof), we would sometimes joke about the amount of global developer man-hours spent on doing so (millions probably if you think about it).

The drive for legacy IE support took a chunk out of the productivity of so many projects.

userbinator3• 5 years ago

Now we have a monopoly of a much bigger, more user-hostile browser(s) and the web "standards" effectively under the control of a single company, which also happens to own YouTube and does things like https://news.ycombinator.com/item?id=17606027 .

This article is actually a nice reminder that YouTube used to work in browsers as "unstandard" as IE6, and with far less processing power than computers today, was not noticeably slow either.

As much as you hate IE6, I hate the current situation far more. Those waging the browser-wars and evangelising the "push the web forward" propaganda have done nothing but contribute to the creation of a different monopoly and the waste of countless processor-hours of time and energy. My personal fuckings go out to everyone who thinks this is "progress" or in any way a good thing.

scoopertrooper2• 5 years ago

For science, I just tried loading YouTube in both Chrome and Firefox. I'd say Chrome may have been marginally faster, but you'd really have to get a profiling tool out to quantify it.

As to the rest of it, not really sure what you're talking about.

userbinator3• 5 years ago

YouTube and plenty of other sites worked fine on the typical computer of the time which would have IE6.

Over a decade later, hardware is much more powerful yet the site is even slower than it used to be. I've started using Invidious exclusively because of that.

Doing less with more is not "progress".

laumars3• 5 years ago

Now we have a monopoly of a much bigger, more user-hostile browser(s) and the web "standards" effectively under the control of a single company

I’m not saying thing are great now but it was definitely worse in the IE6 days. Back then it was a literal monopoly. At least currently we do still have WebKit (Safari) and Gecko (Firefox) plus Blink is open source, which Trident (Internet Explorer) never was.

Better yet, the alternatives these days do still work 99% of the time. Whereas in the early days of Firefox and Opera, and in the later days of Navigator/Communicator, you basically had to accept that a disappointingly high number of sites didn’t render properly at all.

In an ideal world we would see Chrome’s market share much lower, but I’d still take the current ecosystem over the IE6 days in a heartbeat.

Mikeb85• 5 years ago

Google isn't really a monopoly though. They out-compete others for their market share, Chrome is open source, they abide by and push open standards, etc... They don't abuse their position like Microsoft did.

Those waging the browser-wars and evangelising the "push the web forward" propaganda

Really? Thanks to this 'propaganda' the vast majority of software (ie. web apps) runs on every platform there is; Linux, BSD, Windows, OSX, Android, and iOS, kinda on the last one as Apple is dragging their feet with Safari.

trulyme2• 5 years ago

Google isn't really a monopoly though. They out-compete others for their market share, Chrome is open source, they abide by and push open standards, etc... They don't abuse their position like Microsoft did.

You must be kidding, right? If you are not, please take a look at performance problems various Google websites have in Firefox. Or how recaptcha behaves in FF. This is exactly the sort of abuse of power that MS yielded in the days of IE supremacy.

EDIT: about propaganda, Chrome was late to the party and while it sped up the IE demise, it would have happened anyway. Except we wouldn’t land in "all your data belongs to us" browser monoculture in which we are now.

reilly3000• 5 years ago

A lot of folks have wondered why IE 6 stuck around so long in the first place. It’s pretty simple: group policy. Through Active Directory, admins could do things like force a specific homepage, set proxies for all browsers in the company, disable media features for security, make history undeletable, or even turn off right clicking. Around that time there weren’t alternatives to set these policies in other browsers, so admins just disabled the ability to install them in order to maintain tight control. Now Chrome and possibly other browsers can do similar, but less draconian policies, and security concerns are mitigated with separate systems.

After the time this article references, there started to be big changes around how corporate IT was managed. It went from being about gear and gate-keeping to strategically supporting business initiatives. I wonder if this little banner on the YouTube player helped nudge things forward.

Joeri3• 5 years ago

I was stuck supporting IE6 for a lot of large organizations and was often in the loop on why they decided to stick with it. I never heard the ability to set strict group policies as a reason. What I mostly heard was internal web apps built for IE6 which needed to be updated to support IE8, and the incredible pain to actually get those update projects approved. There was also the one customer, a large bank, who had built a custom browser around the IE6 engine, and mandated its use for bureaucratic reasons.

I mostly solved our IE6 problem by moving over our web product to a single page architecture based on ExtJS in 2007. ExtJS had a browser abstraction layer, and as long as you built your screens out of its components you rarely had issues. The major pain points that remained were memory leaks (IE6 was a digital sieve, and SPA made it worse), and performance.

gjvc• 5 years ago

In 2007, I'm guessing that would have been ExtJS 3.x? ExtJS can be a pig to use, but the results are good, in my experience. What would you use nowadays? Have you found anything else with as complete a widget set than ExtJS?

spapas82• 5 years ago

Another thing was that a lot of corporate apps supported IE6 only.

I remember that I was working in an app back in 2010 that was heavily using the showModalDialog https://www.tjvantoll.com/2012/05/02/showmodaldialog-what-it...

This API was not supported by any other browser but was fundamental for our work because you could display a web page in a modal window that would return a value to the parent page when the modal was closed!

It wouldn't surprise me if people are still running IE6 in some virtual environment to be able to access such apps.

hadlock• 5 years ago

Yeah corporate apps had support for IE6 ...and that was it. If you weren't using the internal business tool with IE6, the developers were under no obligation to even respond to your email.

Another big thing was IE6 was supported by pretty much every online bank, particularly in South Korea, which is not a huge market, but SK was one of the last, most advanced economies to switch off IE6, primarily due to banking platforms.

williamdclt• 5 years ago

It wouldn't surprise me if people are still running IE6 in some virtual environment to be able to access such apps.

I've worked on a project ~2y ago to rewrite the software powering the whole core business of a huge company (multi-hundreds-of-millions-a-year revenue I'd guess), the existing software being done somewhat like this, although it was a heavy client rather than web.

A guy developped software to automate his job 25y+ ago with some microsoft tech (FoxPro), somehow the whole company ended up using it and the software kept growing until it was fundamental to the business. For some reason it only ran on something like Windows 95, so they got Windows95 servers that you had to connect to remotely just to use the software. It was incredibly slow of course, and you couldn't copy-paste through the virtualisation layer, which drove everybody crazy for years.

The guy had developped that in his free time apparently (wasn't employed as a software developer), pretty good deal for the company who exploited that for decades and became a major player in the industry.

zxcvbn4038• 5 years ago

Even at places that supported other browsers, it used to be that once every couple years they would pick a new browser version, they would have an entire team spend 18 months testing it on all the applications in use, then roll that out and after a break for Christmas start again.

That blown got blown away completely when Chrome and Firefox started doing updates. People finally got the message that instead of for coding to browsers you code to standards.

The loser was IE because 1) they had been really loose with standards in feeble attempts to lock people in 2) they had stagnated for years while confident their market position could never be overturned 3) by stoically carrying the line that IE was actually part of the operating system and couldn’t be updated independently they locked themselves into a place where IE only got major updates along with major OS releases 4) they had a mess of their double diamond titanium partner program members that would howl every time they tried to fix anything in IE because they had internal applications that depended on the broken behavior in IE and 5) BYOD happened and IT depts no longer had authoritarian control over browsers and their versions.

It worked out for the better I think. Today IE is only used to download Chrome or Firefox though I get a taste of the IE 6 experience trying to find a security profile for IE that actually allows me to hit a URL on the public internet, but it’s all good once you get one of the installers downloaded. Set the default browser to something else and never have to deal with IE again.

TedDoesntTalk3• 5 years ago

Set the default browser to something else and never have to deal with IE again.

Edge could become, hands-down, the best browser in the world. It could automatically put free cake, pizza, and beer into my mouth. And i still would not use it because of the pain instilled on me by Microsoft by IE 6.

8ytecoder• 5 years ago

People are forgetting the history before when IE rose to the peak. It was the exact same story as Chrome. IE did play fast and loose with standards but not to lock you in (may be that as well). Thought like it was to offer better features.

Then the standards improved and the entire IE team was either caught flat footed or possibly even disbanded to work on other projects. IE7 was almost half a decade after IE6

laumars3• 5 years ago

You could still apply Group Policy’s to later versions of IE as well. So that wasn’t the reason. It was more to do with corporate apps being tied to specific versions of IE due to later versions of IE having breaking changes in the way it rendered content vs earlier versions of IE.

Those of us who remember web development in the 90s and early 00s will vouch for how the pain wasn’t just supporting different rendering engines, but even different versions of the same rendering engine.

dejongh• 5 years ago

Agree. Early web development was a compability nightmare. Now that one platfrom rules the world it is less of a problem. It seems that making a standard failed, and we solved the problem by adopting the same platform.

marcosdumay• 5 years ago

So, they kept the browser with the worst security record in history, where one page could own your entire system, because it gives them the means to manage (some of) its security?

Ok, looks like business as usual.

noduerme3• 5 years ago

From the article... >>Once they realized what had happened, they cornered our boss for details, grappled with the consequences of our actions and begrudgingly arrived at the conclusion that the ends had justified the means

I totally applaud the guerrilla methods taken by employees, unbeknownst to the bosses, that are described in this essay. Sometimes you have to just cut out the committee. Getting people to ditch their ancient software so you can move on with features is virtually impossible unless you hold a gun to their head, and that is never something tasteful for the higher brass -- even the ones who demand the features in the first place. They're in the miserable position of keeping every single person satisfied and trying to improve things. That's why it's not terribly surprising that they're happy to have the engineers go rogue every once in awhile.

cube005• 5 years ago

If you have a flat management structure sure, but in this case they knew they were breaking the company's change management procedures because they were going to "slip the code into production without anyone catching on"

I'll wager they were strongly warned they might not be so lucky if they try a stunt like this again. If something had gone wrong they'd all be out of a job and good luck getting a favorable reference.

noduerme3• 5 years ago

So? You do realize that most people who write code can easily find other jobs, and are usually paid better for thinking outside the box, right?

>if they try a stunt like this again

Dear Principal Skinner - I will never make management look like idiots again. Sincerely, Bart.

ajkjk• 5 years ago

I mean... who cares about the reference, these people are gonna be imminently employable anywhere. The only thing on the line for them was _this_ job, which they probably liked, but doing a good thing and risking something you like is just courage.

cube005• 5 years ago

Next, we just needed a way to slip the code into production without anyone catching on.

The time when I need to "slip the code into production without anyone catching on" is the time I know I need to move on from a company.

This isn't an action movie, it's a business, the business has procedures, if you don't like them, either try and change them or leave, you should never go rouge like this because if you don't pull it off flawlessly you will be terminated.

Even if you succeed, it will be recorded in your file and wheeled out if they ever want to turf you for any other reason they feel like in the future because now you have a "history"

sangnoir2• 5 years ago

you should never go rouge like this because if you don't pull it off flawlessly you will be terminated

Live a little; they knew the risks going in and were convinced it was worth getting terminated. I have staked my employment on things I strongly believed in before. The corollary was I wouldn't want the job anyway, if the organization disagreed.

tardyp• 5 years ago

Good ol times when front page videos were 40k views max...

gpvos• 5 years ago

This is probably the best example ever that it's easier to ask forgiveness than permission.

axegon_• 5 years ago

Fun fact: I know a frontend dev who only used ie6 for development for as long as I knew him which sounded like masochism. But he had a point: "hey, if it works on ie6, it will work on a potato with a VGA stuck in it".

TimTheTinker• 5 years ago

That's a badly misguided take though. I bet his apps don't work anymore now.

Lio2• 5 years ago

“It’s easier to beg forgiveness than ask permission”.

Never truer.

marcodiego• 5 years ago

TLDR: "The plan was very simple. We would put a small banner above the video player that would only show up for IE6 users. It would read “We will be phasing out support for your browser soon. Please upgrade to one of these more modern browsers.” Next to the text would be links to the current versions of the major browsers, including Chrome, Firefox, IE8 and eventually, Opera. The text was intentionally vague and the timeline left completely undefined. We hoped that it was threatening enough to motivate end users to upgrade without forcing us to commit to any actual deprecation plan. Users would have the ability to close out this warning if they wanted to ignore it or deal with it later. 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. We even delayed having the text translated for international users so that a translator asking for additional context could not inadvertently surface what we were doing. Next, we just needed a way to slip the code into production without anyone catching on."

praptak• 5 years ago

Compare this to how Google handled Kathryn Spiers for unauthorized posting of an internal banner.

We can at least stop pretending it was for abuse of commit rights and not for labor activism.

captainredbeard• 5 years ago

Labor activism isn’t a job role

seoaeu• 5 years ago

It is however protected under the National Labor Relations Act: firing someone for engaging in concerted activity is a violation of their rights

forgotmypw17• 5 years ago

I think any behavior trying to influence the user's browser CHOICE is incredibly patronizing and self-serving.

Web sites should bend over backwards to support whatever the user happens to have, but everyone is just too fucking lazy.

fusioneventz• 5 years ago

Thanks for sharing < a herf="https://fusioneventz.in"> Niranjan </a>

samgranieri• 5 years ago

Thank you for doing this. I was given photoshops for a hobby site I maintained and IE6 was always foiling my plants. I was working as a contractor for Tribune Interactive in the fall of 2009 and I heard stories about how the senior front end developer almost got in a screaming match with high level execs demanding they upgrade their browsers. It worked

A Conspiracy to Kill IE6 (2019)

Related Stories