Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

A conspiracy to kill IE6 (2019)

blog.chriszacharias.com

Post reply on HN
Print view View original

31–40 of 113 posts

Re: A conspiracy to kill IE6 (2019)

#31

Post by ksherlock »

cogman10 wrote: post #4

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 rende…

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.

Re: A conspiracy to kill IE6 (2019)

#32

Post by DaiPlusPlus »

jwells89 wrote: post #21
bitvoid wrote: post #13

Earlier quoted context omitted.

It was forked over 10 years ago now. I doubt their shared past is even recognizable now.

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...)

Re: A conspiracy to kill IE6 (2019)

#33

Post by zx8080 »

jwells89 wrote: post #21
bitvoid wrote: post #13

Earlier quoted context omitted.

It was forked over 10 years ago now. I doubt their shared past is even recognizable now.

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.

> whereas the Blink team wanted it to be part of Chromium with non-Chromium users of Blink needing to bring their own multiprocess.

Is that why Electron basically means Chromium?

Re: A conspiracy to kill IE6 (2019)

#34

Post by jwells89 »

DaiPlusPlus wrote: post #32
jwells89 wrote: post #21

Earlier quoted context omitted.

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.

Re: A conspiracy to kill IE6 (2019)

#35

Post by summerlight »

rekttrader wrote: post #7

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.

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].

[1] https://web.archive.org/web/20141217174028/https://www.moder...

Re: A conspiracy to kill IE6 (2019)

#36

Post by Macha »

jwells89 wrote: post #34
DaiPlusPlus wrote: post #32

Earlier quoted context omitted.

> 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.

Nah, QtWebEngine is using Chromium, not Blink directly

Re: A conspiracy to kill IE6 (2019)

#37

Post by tracker1 »

cogman10 wrote: post #4

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 rende…

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.

Re: A conspiracy to kill IE6 (2019)

#38

Post by jwells89 »

ksherlock wrote: post #31
cogman10 wrote: post #4

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 rende…

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 prob…

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.

Re: A conspiracy to kill IE6 (2019)

#39

Post by chupchap »

Octoth0rpe wrote: post #17

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.

As a student I would struggle to put out a page that would render the same way in all browsers. It takes so little effort now and I'm so glad for that

Re: A conspiracy to kill IE6 (2019)

#40

Post by ttfkam »

ksherlock wrote: post #31
cogman10 wrote: post #4

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 rende…

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 prob…

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 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.

Post reply on HN