One of the best things that's happened to the web is all browsers moving over to an evergreen model of distribution. IE was such a cancer on the internet, everyone in the bad old days programmed to the lowest common denominator (IE6, 7, or 8). Even while brand new features and standards were widely adopted by everyone else.
One of the worst things that's happened is nearly everyone converging on using webkit to render everything. Feels like slowly stepping backwards into the days when everyone writes for webkit rather than a more open internet.
That being said, there's so little difference between webkit and gecko now-a-days that it's really hardly a major issue.
The thing about IE, one thing anyway, is that it's CSS support, box model, etc, was broken and non standard. So you would need to exploit CSS parser bugs to give IE a separate CSS, use DXImageTransformFilter to make png transparency work, etc.
Maybe by WebKit you mean Blink and by Blink you mean V8, because when somebody does a "Show HN: something that doesn't work in Safari or FireFox" it's usually a javascript problem not a CSS or rendering problem. And every 6 months someone complains that Safari is literally worse than IE 5 because it doesn't support some non-standard privacy nightmare API that Google added last week.
I came to hate IE with every fiber of my being, but let's call a spade a spade here: IE basically invented modern CSS. Netscape 3/4 and IE 3/4 had very different views on how web pages should be dynamic and how they should be styled.
IE championed CSS and the DOM. Netscape was fixated on bgcolor, font tags, etc. Netscape introduced the <layer> tag. IE just made every tag a potential layer. Remember AJAX? Loading content from a script and injecting it into a page? That was 100% IE.
The issue wasn't that IE 5, 5.5, and 6 were bad when they came out. They weren't. Not one bit. They were absolutely the cutting edge. Then Netscape died, Microsoft had no more competition in the field, and so they basically just stopped. For years.
Mozilla (and shortly after, Firefox) prompted an eventual response, which was IE7: I minor update to an otherwise abandoned codebase. Once it was clear folks were seriously looking at dumping IE, IE8 followed shortly after. The CSS Acid Test showed how far IE had fallen behind, so IE9 and IE10 moved to stop the bleeding. IE 11 was the final release before Microsoft finally dropped the legacy IE baggage, aka the Trident engine. By that time Chrome had stormed the field, and Edge's renderer just couldn't keep up relative to the amount of money Microsoft was willing to pay. So they dumped their own engine and moved to Blink. (If you can't beat em…)
But let's be clear, between 1998 and 2003, IE was at the forefront. Its CSS support wasn't broken. It was the best and only. Its box model wasn't broken. There were no other box models to compare it to. You didn't need to give it separate CSS, because it was the only one around really to parse it. It didn't have non-standard AJAX support. It WAS the standard since no XMLHttpRequest object existed at the time.
You could write VBScript on a web page instead of JavaScript and have a >90% chance that client could run it. Flash didn't have to be bundled or installed separately, ActiveX just took care of it in the middle of your browsing session.
It was the Bronze Age of the web. IE was the one that moved us beyond stone tools. Later supplanted by Firefox and Safari's Iron Age, but still an important and arguably necessary step toward the modern web we know today, including the box model.
What was bad about IE, more than anything else, was how tightly integrated to the OS it was. Running windows 2000? You can only run IE6. Running windows XP? IE8 is your limit.
This was particularly problematic because MS struggled MIGHTILY to get Vista out the door. It was a decade of really terrible IE which was pushed out longer due to Vista's poor performance and stability.
Meanwhile, you could install firefox and later chrome and get a relatively modern internet browsing experience.
Ironically, I think the thing that hurt firefox adoption was the switch to the evergreen model. The era of firefox 3->~33 was quiet painful, updating was simply far too intrusive. "You are out of date, click here to download the latest version!". Chrome from the get go was pretty good at making that seamless.
My memory may be failing me but I think Firefox 3 was also where it started to feel less like the snappy minimal utilitarian “just a browser” it was in versions prior and started to feel more clunky, branded, and product-y, which also didn’t help its adoption in my opinion.
If nothing else it primed users for adopting Chrome, which originally had a less branded, more utilitarian feel not unlike early Firefox.
This cycle is somewhat repeating itself though as Chrome has become very branded and product-y. Would love to see something similarly functionally-minded as early FF/Chrome come out and take a big bite out of Chrome’s marketshare.
I think the update experience was a small part of it, but for me the process model of Chrome was the main thing that made me eventually switch. Crappy javascript on one tab would lock up the entire browser -- a thing that just simply doesn't happen in Chrome.
> IE championed CSS and the DOM. Netscape was fixated on bgcolor, font tags, etc.
Netscape championed its own alternate to CSS: JSSS. Javascript-based Style Sheets.[1] But the W3C went for CSS, so the CSS implementation in Netscape 4 (1997) was very rushed, and poor quality.
This is what led to the @import hack being possible, which used different style sheets for IE and Netscape.[2][3]
Yup. When it was released, IE6 was the thing. It was fast and worked very well. Speaking of bronze age, people forget how slow js was then. I remember writing server code that generated js code of pre-populated arrays to push down to the client to use for interactivity. Sending json and populating on the client was slow enough I could count in my head while watching the screen. Array look ups though were fast enough.
> But let's be clear, between 1998 and 2003, IE was at the forefront. Its CSS support wasn't broken. It was the best and only. Its box model wasn't broken. There were no other box models to compare it to. You didn't need to give it separate CSS, because it was the only one around really to parse it. It didn't have non-standard AJAX support. It WAS the standard since no XMLHttpRequest object existed at the time.
They should have done what Chrome does - write the standards to match what you've implemented.
You're completely correct that IE 4.0 - 6.0 was kicking Netscape's butt in terms of CSS support (not difficult as Netscape 4.x was truly awful). But not quite right that it was the best or only until 2003.
Opera could probably claim best CSS in the late 90s. And by 2003, Netscape 6.0 & 7.0 and Mozilla Suite 1.0 would have better CSS 2 support than IE 6.
I remember those days well. IE was really the only usable browser for a long time. Netscape was incredibly unstable, didn't support any modern standards. Mozilla (before Firefox) was too new, also unstable until roughly 2005.
> Mozilla (before Firefox) was too new, also unstable until roughly 2005.
That doesn't match my memory at all, unless you include Phoenix under the Firefox umbrella. I switched … I don't remember exactly when, probably in late 2003, but certainly when it was called Phoenix, and it was such a breath of fresh air coming from IE. I'm sure there were breakages, because it was the early noughties and that's what the internet was like, but I don't remember a single time wishing for something that IE provided that Phoenix didn't.
But I was just coming to it as a user, not as a dev—maybe you're referring to the dev perspective.
I was including all "Mozilla" browsers (Phoenix, Firebird, whatever) under that umbrella. I didn't start using Firefox seriously until roughly 1.5. Maybe it was just me!
I was there too. Actually ever since Mozilla M1 or M2 I think. Things were pretty rough back then (so much blue!). Then the browser-only Phoenix followed by the rename.
I (and presumably others) hated the notion of a Microsoft/Windows-only web. Younger folks just don't realize how close we came and the debt owed to those Mozilla devs.
And maybe peaked at 3% marketshare between 2000 and 2003 during Apple's early resurgence. Remember that Apple almost died in the late 90s. It wasn't the juggernaut it is today by a long shot.
Mac IE became its own albatross for web development for a while once Safari hit its stride and supplanted it.
Prior to somewhere between 2003 and 2005, "web standards" were aspirational rather than the reality on the ground. There was Internet Explorer for Windows at 85-90% and all the bit players scrambling for a percentage point or two.
For the bit players, the standards were their only bid for survival since no one would tolerate putting extra dev time into an extra 1%. Together, you could make a case for a 10-15% target.
For Microsoft's part, "IE is the standard" had an extremely long tail.
That DirectX filter hack for PNG transparency haunts me.
What makes it so much worse is that Tasman, the engine used in IE for Mac, had awesome support for CSS, transparent PNGs, and more years earlier, but for some reason it got dumpstered when IE for Mac was discontinued while the inferior Trident lived on in Windows.
Microsoft did a lot of weird things like that - also with Outlook, which used to use Trident to render emails, when IE was finally getting a bit better the Office team decided to switch Outlook to use the far, far worse Word HTML rendering engine instead...
Even thinking of that makes me very happy I haven't had to worry about formatting email newsletters for more than 10 years now, brings back bad memories...
Safari has been a huge pain for me as a web dev for its nonstandard parsing/implementation of basic CSS and requiring special cases made for it for things like modal focus/scrolling, nothing to do with Google or privacy whatsoever.
> The thing about IE, one thing anyway, is that it's CSS support, box model, etc, was broken and non standard.
Box model in particular was because they used the much saner border-box instead of the official spec's content-box, which was what everyone else used. You could make the other browsers act like IE by specifying it with box-sizing in the CSS, or from IE8 onwards you could make IE act like the others.
> One of the worst things that's happened is nearly everyone converging on using webkit to render everything. Feels like slowly stepping backwards into the days when everyone writes for webkit rather than a more open internet.
There were substantial differences even from day one. As I recall one of the differences of opinion that was a driving cause of the fork was the multiprocess model; WebKit team wanted that to be part of the framework, letting any WebKit-embedding app get it for “free”, whereas the Blink team wanted it to be part of Chromium with non-Chromium users of Blink needing to bring their own multiprocess.
> with non-Chromium users of Blink needing to bring their own multiprocess.
Isn't that what happened anyway? The non-Chromium dists of Blink like Electron and WebView2 both employ the same multi-process approach as Chrome.exe (and my RAM availability is suffering for it...)
Electron bundles Chromium, not Blink directly. Believe this is true of WebView2 too but am not well-versed in Windows internals.
In fact I’m not aware of any major non-Chromium usages of Blink, except maybe in Qt (where it replaced WebKit a few years ago) but I’m not sure about that either.
Webkit isn't the issue. The issue is Apple is making the same mistake as Microsoft by tying a browser version to a version of the OS. With no upgrade path for people with older phone, you have support the lowest common dominator instead of pushing people to install a new browser.
OTOH, Apple does tend to get OS updates to their users a lot faster than anyone else with market share. At least until they stop supporting a beloved device.
Browser engine for me is like a Linux kernel for me.
Everyone should be banding together to work on one. Not making a separate one each year. We have js libraries and frameworks for that. Of course some innovations that requires to be done in engine won’t happen as fast - but on the other hand amount of innovation in software that runs in browsers will be much higher.
We saw that with IE “innovations” of Microsoft blocked decades of innovations in web apps.
If I go to its Wikipedia page Blink is backed by many more companies than just Google. Where WebKit is or was basically Apple and some companies. Not dropping links because anyone can check wiki pages.
Those companies listed on Wikipedia are ones which have contributed over time, they are far from exhaustive and lack a lot of context.
WebKit itself was a fork of KHTML by Apple, and Blink was a fork of WebKit by Google. Both initially started with the one company and then had others contribute to the open source projects.
In the WebKit wiki it is noted that WebKit had 50% market share at its peak. This is when IE and Firefox still had about 15-20% market share each. WebKit was well on its way to being where Blink is today. It was used by Apple, Google, Palm, Sony, Nokia, Nintendo, and more.
It wasn’t the industry banding together to create Blink. It was Google. Then one by one, various companies gave up the fight, Microsoft and Opera are two of the most notable.
Maybe if Apple was easier to work with Blink wouldn’t have happened, but I probably had more to do with Google having their own view of what the web should be, and their own engine lets them push that on everyone.
The difference with today and blink is that it's open source and if there's a serious issue it will be forked. And has been, twice, in order to get to Blink.
The reason it's not a major issue anymore is that most developers are smart enough to follow the web standards first and then make sure the browser supports it. They've found that following standards--just like all other programming--is the foundation of computer programming.
I don't find it okay to bash IE so harshly. Sure it wasn't the best but no one forced people to use it.
And I don't think the conspiracy in this article is something to boast about. It's literally using dishonesty with the higher-ups (no matter how much you might hate them) as a way to make your own job easier.
Plus, stuff like this is why we ended up in a Chrome monopoly, to the point that even Edge finally got Chromified.
In the early 2000's, IE was really the only viable option. Nobody else provided a fully functional browser. Netscape was rotting. Safari, Firefox, Chrome were not yet released. If you wanted something that worked and didn't crash every 30 minutes, IE was it.
> Sure it wasn't the best but no one forced people to use it.
Ehh, plenty of corporate IT decided to force employees to use IE. That's actually my main interaction with programming for it, clients that HAD to use IE due to corporate policy.
> I don't find it okay to bash IE so harshly. Sure it wasn't the best but no one forced people to use it.
No one was forced to use IE, but many people had the decision made for them. Sometimes it was because it shipped as the default on the dominant operating system. (I suspect people forget how dominant Windows was. While Windows remains the dominant operating system on desktop computers to this day, their share wasn't diluted by platforms like Android and iOS.) Other times it was because sites designed exclusively for IE were visibly broken in other browsers. (I suspect part of the reason why I am not bothered by incompatibilities with Firefox to this day is because the incompatibilities are minor compared to those we faced in the past.)
I do agree with your sentiment about conspiracy boasting, though from a different perspective. Their actions represent a group of people in a position of power making a decision for other people, even if it wasn't an officially sanctioned decision. (Though I have no doubt that many people cheered for their actions.)
I and every single other webdev I've known from the early aughts would immediately offer to buy this person a beer upon hearing this story in person. Those claiming that safari is the new chrome just don't understand how bad IE was at the time.
> Those claiming that safari is the new chrome just don't understand how bad IE was at the time.
I suffered in the IE6 times, understood how bad it is, but still claim otherwise; when we're saying that Safari is the new IE/the new Chrome, it's with good reason. We've recognized that a piece of history is repeating itself, and it's pretty sad that it gets handwaved away or dismissed, just because people happen to have bought into its ecosystem.
> just because people happen to have bought into its ecosystem.
I don't think that's a fair take. Plenty of us are developing webapps for safari/chrome/firefox, are barely ever consider browser differences. I think a lot of credit for this also goes to a wealth of high quality css/js frameworks that abstract what minor differences exist these days, rather than apple.
It's worth also considering what specifically was bad about the IE dominance era in order to understand how it maps to our current situation. IMO it wasn't just that IE was bad (of course it was), but that its usage share was _dominant_. That doesn't come close to reflecting the current state of mobile browser usage: https://gs.statcounter.com/browser-market-share/mobile/world...
Worldwide Safari is a minority player. In the US (probably the best case), safari holds a slight majority at 55%, and the other 45% are likely almost entirely blink. At one point (early NS6/pre-phoenix), IE had something like 95% usage, and closer to 99% for corporate usage.
It's still bad. The awful continues for end-users.
Just this afternoon, I witnessed one elder tech teaching another elder tech how to enable IE mode on Windows 11 because that is the only way to get a particular hardware-based NVR to be viewable.
This is a red flag that you are trying to use non standard CSS or you are trying to use things that are so bleeding edge that you are requiring even evergreen browsers to have been restarted and updated very recently. Set your target to any CSS that has been around for about a year or so and is based on standards and you shouldn't need to be writing anything different for specific browsers.
I haven't had to target a specific browser for any CSS in years. I honestly don't know what people are trying to do that would require this now days. Grids, flexbox, animations, transitions no issues and don't even need browser prefixes. The only thing that I can think of that still requires a browser prefix in safari is -webkit-backdrop-filter. Maybe I'm just not doing super out there stuff I guess.
Netscape 4 was especially bad. Trying to do anything interesting with CSS was playing with fire -- sometimes it'd work, sometimes it'd break, sometimes it'd crash the browser entirely, and there was no way to tell which one you'd get.
I worked on the front end of CollegeHumor.com at this time and spent a lot of working hours bashing my head against the wall debugging for IE6 (and even IE5).
I don’t remember the decision/approval process that happening internally but we put up a similar banner very shortly after YouTube added theirs. We continued supporting IE6 for some time but started prioritizing it less and less.
No other software I hated more than IE6 in my life. I've started webdev in 2003 and was doing PSD to HTML/CSS conversions in the evenings and nights after my webdev dayjob. Opera and Firefox was a pleasure and easy to convert for, IE6 caused no end of pain for these conversions. In time I've learned enough tricks to know all the workarounds and could look at a given PSD file and know within 30min precision how many hours it'll take me, at that point I've started charging clients on top for IE6 support, this was around 2008.
This IE6 hate transcended into me avoiding all MS products for a long time. Only after they bought github and linkedin I gave in.
I was still messing around with IE6 as late as 2012...the web agency I worked at had clients in huge old school companies who must have had proprietary internal software that depended on it. Whenever we thought we could finally ditch it, we'd get forwarded an email from a higher up at a client company with a subject like "??????" containing a screenshot of our pre-launch environment for a new site looking entirely broken in IE6.
At that point we had this giant wad (like, several files and multiple megabytes?) of JS and CSS of uncertain provenance that we'd conditionally load for IE6 that magically made almost all of the more modern stuff actually just work. I'll never forget the feeling of rage-then-relief when the screenshot situation above happened a matter of hours before a major site launch. I booted my IE6 VM, shrieked at how utterly broken the site was, copied over that ungodly blob of polyfills and dark magic, reloaded the page and everything looked correct. I was working in a coffee shop and may have whooped out loud.
On a tangent: the screenshot of the deprecation banner brings back memories of a time when the internet still felt relatively free, before the rise and dominance of mega platforms in the 2010s.
I still mourn the loss of community forums that were replaced by reddit, or worse, Discord. So many hours spent on various phpBB/vBulletin etc. forums. :^)
I run a Discord and a phpBB forum. The Discord is free. The forum is $130 a year and signups are broken. It's very clear why everyone moved away from them.
>The freedom to run the program as you wish, for any purpose (freedom 0).
That's pretty wild that a direct competitor targeted a browser based on a few random engineers's personal views. In 2024, privacy being front and center, I wonder if its even feasible to kill off Google. They are slowly becoming a bully and a menace to society (unless you're one of the three letter agencies).
They targeted IE 6 or older by recommending folks install IE 8, Firefox 3, or Chrome 4, shown in random order. Since Firefox and Chrome were self-updating, they didn't have the problem of obsolete versions floating around.
Microsoft was already offering users to upgrade to IE8. Users chose IE6 for whatever reason. Most likely because non-technical users didn't give a shit, or maybe they were stuck on IE6 due to ActiveX or other compatibility issues. Google faked a "we will be phasing out support" screen to trick users into changing the browser. All this talk of open source and "open" this and "open" that, and yet they still want to manipulate the user. How about you let the users choose? And if they choose wrong so be it. I guess it is understandable given they are in the advertising business. Maybe Google has others' fooled, but they haven't fooled everyone yet.
By your definition Hacker News would be anti-free because NCSA Mosaic 1.0 can't render it. Youtube didn't add anything to block IE6, they simply removed all the non-compliant hacks that barely kept IE6 working as long as it did. IE6 can't render a website of purely standards compliant code. They didn't fake anything. If you went to youtube.com on March 13 2010, the day they deleted the IE6 helper hacks, the videos wouldn't play.
>IE6 can't render a website of purely standards compliant code
Yes, supporting IE6 is a burden, so just stop supporting it. A simple "IE6 support ending on Mar 13, 2010. Please upgrade to IE8 to continue using this website." would suffice. But they tried to trick the user into installing a random browser through a fake banner.
"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."
Google doesn't give a shit about the user anyway, they are anti-standards unless it helps them collect user data or show ads.
All their primary products are proprietary and they don't support open data exchange. This in and of itself isn't a problem to me (if a company is open about their business model) but the smugness of the company and the employees with their "we know better so we are going to trick or force you" is repulsive.
Not saying IE6 is good. It's quite comfortable to have the left hand owning the Chrome and the right one blocking other browsers with very specific changes in one of the most popular web sites. Helps power, control, and ads, you know.
The similar thing is happening now [0] (however the reasons are showing ads - but essentially they've always been to show ads!). But this Firefox cow is too precious to kill it entirely: it prevents antimonopoly investigation as "see there are others" on the field of browsers.
Back when I worked on websites (2011-2012), supporting IE6 through IE9 ate somewhere between 1/3 and 1/2 of all my development time, as nothing was consistent with any of them, and that included the same browser with slight tweaks. It was a horrible, time-wasting badly behaved set of browsers and because I worked on government sites for part of that, Internet Explorer support was mandatory.
I'm so very glad they're gone - they were the lemon cars of the computer world of those days.
"A small team of [hero] web developers conspired to kill IE6 from inside YouTube and got away with it [saving us all from eternal wackness]. - There, I fixed it for you. :)
The thing that gets me the most is how ancient 10 year-old YouTube looks. Which makes sense that this was the era before smartphones, when IE6 still had significant market share. But 10 years! Sounds like such a short time.
Actually, it was 15 years ago now. The article says "ten years ago" but it was posted in 2019, so 15. I freaked out too when looking at that really old UI and thinking "that was in 2014?" but nope
Killing Safari would hand control of the internet over to Google. This requires everyone trust that Google will do the right thing now and forever, and not work in their own self-interest. That’s not a bet I’m willing to make.
So what you’re saying is you helped your billion dollar mega-corp use their monopoly position to destroy another monopoly and ensure chromium would be the next IE.
That's a pretty wild take, considering that the author specifically said they randomized the order of suggested browsers. They didn't push Chrome, they pushed "anything but IE6".
The grandparent comment would only be somewhat revelant in future versions of Youtube and Google services, since they did end up pushing users to 'try' Chrome exclusively (for Youtube in the footer and everything else via top banner/popups, for both Firefox and even non-Chrome Chromium browsers).
Which given the lawyers' concerns in this article is curious the 180 they did.
I distinctly remember at the time that Google even had a browser extension that loaded chromium inside the IE6 viewport...
Shady browser advertising toolbar companies were using that extension to hijack and silently upgrade users into Chrome. Google bought that company as well.
It's exactly the opposite. At the moment, Microsoft had a really hard time to move users to IE7 mainly because their distribution model does not allow automatic upgrade at the time. You need to go to MSDN, download IE7 then manually install it. It's pretty clear MS was happy with this deprecation given that their response was a website called ie6countdown[1].
Browsers have a tendency to consolidate. I think "there is a dominant browser used by N-times-% of the userbase relative to the second-most-popular" is a feature endemic to the ecosystem.
Much like "You can switch off Google search any time you want, but people will keep using it as the default until there's another clear better option, and then most people will switch to that instead." If browsers are pretty interchangeable, why not use the best?
One of the worst things that's happened is nearly everyone converging on using webkit to render everything. Feels like slowly stepping backwards into the days when everyone writes for webkit rather than a more open internet.
That being said, there's so little difference between webkit and gecko now-a-days that it's really hardly a major issue.