In the enterprise space, IE6 usage has fallen significantly from Dec 2010 to Dec 2011 according to my own observations, but it's still not near enough zero for my own comfort:
Dec 2009: IE 6.0 share is 37.3%
Dec 2010: IE 6.0 share is 15.5%
Dec 2011: IE 6.0 share is 6.8%
and just for the rest of the IE figures from this particular site in the enterprise (sorry, not comfortable releasing actual session counts, all I can say it's in the hundreds of thousands of unique users):
External site, accessed by external companies. Granted, we only "support" the products on IE and Firefox, but the usage is still pretty varied, as they all mostly under other browsers.
It looks like IE6 and IE7 are just being dropped as that organization adopts Windows 7 for new hardware as old hardware gets triaged out whenever it breaks.
Not that it's an entirely bad thing, but since IE8 is becoming more standard, wouldn't updating to IE9 be easier?
Windows XP is still being supported by Microsoft (until April 2014). I think that IE8 is the most recent version of Internet Explorer which will ever work on Windows XP.
Question really is whether supporting it costs less than 1% of your total revenue. John Resig has said that very little of jQuery source is IE6-specific. That is to say, supporting IE7 and IE8 is a pain in the neck, and IE6 comes along for the ride.
Maybe in terms of Javascript, but IE has so many frustrating CSS bugs that aren't as bad in 7/8 that I draw a very clear mental distinction between them.
Another major point is testing & debugging. IE6 had a basically useless JS debugger. For IE7 & 8, you can at least use IE8's developer tools to pinpoint a bug and fix it (IE8 comes with IE7's rendering engine built in). With IE6, you're very often stuck looking at Fiddler dumps or binary-chopping lines out of your programming until you've pinpointed the error.
Back when Google Search still supported IE6 as a first-class citizen, I found that very few lines of code were needed to make IE6 work. However, figuring out what those lines of code were could easily take more time than making it work in every other browser. Just the amount of time lost because I accidentally included a trailing comma made me want to throw my Windows laptop across the room.
I like how ie6countdown.com renders perfectly fine on both IE6 and modern browsers.
I don't like how inaccessible HTML5 and CSS3 techniques are to older browsers. It reminds me of having to upgrade flash.
I like how we try to make sites accessible for the blind, regardless of their usage stats.
I don't like how we seem to throw away 1% of our best possible conversion rates, by fully ignoring 1% of our audiance.
I like how the web is maturing and growing.
I don't like how plain and simple information-providing websites are turning into HTML5 applications, with 100k's of javascript, hashtags and other dynamics.
I believe that in 20 years, sites that were build to render on IE6, will continue to render just fine. Sites that were build using experimental browser vendor-specific code, with AJAX and hashtags might need a special server to render. Is that progress?
EDIT: seems to be some confusion about "Sites build for IE6". I ment "Sites that render on IE6/were build with IE6 in mind". To me that doesn't auto-translate to active-x, MS-filters, conditional rules, IE7.js and CSS hacks, but I can see how others can view that. Anyway, I am clearly playing with fire, by taking these views on IE6. I'll just let this be and not delete it. It wasn't a troll or a flame, but this topic is always a heated one, so best to just let it be.
I'm okay with newer pages not being viewable by browsers more than 10 years old. Upgrading your browser every half decade or so seems eminently reasonably, if only to ensure you get all of the security patches that are no longer being backported to EOL browsers. Of course, with browsers like chrome, they are auto-updating all the time anyways - so it's not like you have much of a choice.
What is more important to me is that older web sites continue to be browseable (to some degree, bit perfect rendering isn't essential) 20 years from now.
Perhaps because I didn't say nor mean both those things. Perhaps also because your parser mistranslates "sites designed for IE6" as "sites using MS-specific code".
>I don't like how inaccessible HTML5 and CSS3 techniques are to older browsers.
Because these technologies came into existence after these browsers were made. Such is progress.
>I like how we try to make sites accessible for the blind
This is a law, and being blind is rarely a choice.
>sites that were build to render on IE6, will continue to render just fine
Because they were built with two codebases, one for regular browsers, one for IE6. Double the work for that 1%.
>experimental browser vendor-specific code
Browser specific code? Yes, if that code doesn't become standard. But IE6 required a lot of browser-specific code. The hopes are that the browser specific code gets the recognition for its usefulness and becomes part of the still-in-progress standard.
>> Because these technologies came into existence after these browsers were made. Such is progress.
An example: the "about page" for Google Music. You need to download the newest Chrome to view it. I don't call it progress if accessibility and backwards compatibility get the shaft.
>>This is a law, and being blind is rarely a choice.
Two things: I hope you make sites accessible for the blind, not because it is law, but simply because you want the blind to access your content. Put another way: If it wasn't for the law, we could forget about the blind too? Accessibility needs a law for front-end engineers to take note?
Being on IE6 is rarely a choice too, especially in a corporate environment.
>>Because they were built with two codebases, one for regular browsers, one for IE6. Double the work for that 1%.
You don't need exceptions to make sites render on IE6. You only need exceptions if you start implementation in a modern browser, and make IE6 a minute after-thought.
>> But IE6 required a lot of browser-specific code.
Only if _you_ required hacks to make a design work.
>>Is that progress? Yes.
I don't think that was the progress that Tim Berners Lee had in mind. A progress where HTML documents from the HTML5 era become inaccessible to later generations.
Disclaimer: I am able to build sites that render in IE6, without much hacks or troubles or extra work, but I don't like working around its quirks. I don't have a strong opinion on why and when IE6 should die, and I certainly don't want to play that judge with my clients websites. I want maximum accessibility, and that includes IE6 in my opinion.
>You don't need exceptions to make sites render on IE6.
I disagree. IE6 doesn't support a lot of standards that other browsers supported, mainly due to not keeping up with changing standards.
>If it wasn't for the law, we could forget about the blind too?
Note that I said "being blind is rarely a choice", implying that we should code for the blind because they did not choose their limitations. If you're forced to use IE6 in a corporate environment, it's because the designer of a web app wrote it for IE6, not for "the internet Tim Berners Lee had in mind". No one is forcing you to use entertainment-based web apps at work, in fact your management might discourage it.
>I don't think that was the progress that Tim Berners Lee had in mind. A progress where HTML documents from the HTML5 era become inaccessible to later generations.
And you think that the designers of floppy disks intended for them to be thrown out when USB drives came into existence? THIS IS PROGRESS This is what progress does. Technology changes. Don't code for current technology, code for maintainability. If your code works on IE6 but is broken in newer browsers, your code is broken by design. There is a standard, and code that follows it will work in browsers that respect it. IE6 does neither.
You honestly seem to be arguing that IE6 was a perfectly compatible browser and was the pinnacle of browser design. I think you'll find a lot of disagreement, even from Microsoft themselves. We know enough about security these days to know that not upgrading is a bad idea.
>> You honestly seem to be arguing that IE6 was a perfectly compatible browser and was the pinnacle of browser design. I think you'll find a lot of disagreement, even from Microsoft themselves. We know enough about security these days to know that not upgrading is a bad idea.
If I honestly seem to be arguing that, I apologize. It also means this discussion is fruitless, because counterarguments are based on a perception that no one here holds. If anything I view IE6 as an accessibility issue.
> You don't need exceptions to make sites render on IE6.
>> I disagree.
Hackernews renders on IE6 and it doesn't need exceptions. You don't need exceptions to make sites render on IE6. You might need a work-around or two (for example: opting for padding, instead of margin). You can still both code to standards and have it render without exceptions on IE6.
>>implying that we should code for the blind because they did not choose their limitations.
I disagree. User choice or forced, disability or no: accessibility is for everyone. Wether people choose to turn off javascript or their browser forces them shouldn't make a difference. Choice is no pre-requisite for accessibility, unless you make that choice for your users.
>> THIS IS PROGRESS This is what progress does.
In your analogy, at least you still have access to the data on the floppies. And if our current progress destroys the initial ideas for progress from authorities like TBL, doesn't that give reason to pause and reconsider the way we are heading?
No offense to TBL, but the Internet is not what he designed anymore. It hasn't been for a long time. He created an idea, and others took that idea and ran with it. To your point, yes sites render in IE6. Hacker News also looks like shit and was poorly put together.
There's something to be said about backwards compatibility. It doesn't exist, it shouldn't exist, and clinging to it holds back progress. Well designed code should just work, full-stop. Look at the jumble that is Windows, or Linux. Mac OSX is highly regarded, it's fast, its stable, and it "just works" so they say. It also has no backwards compatibility to worry about. How many lines of code are in Windows or Linux solely to make antique software work because the designer used unportable code?[1]
Design your production sites to the standard and they will last forever, regardless of what browser you use. Design them for IE6 and they will forever be broken.
Regardless, neither you nor I can change the course of the Internet. All we can do it sit back, enjoy the ride, and write code that works.
"I don't think that was the progress that Tim Berners Lee had in mind."
I don't think he had IE6's non-standard quirks in mind, either. Had they not come up with their own standard, the concept of graceful degradation would certainly have been easier.
It was Windows XP that led to this, and the failed Vista OS launch that took 5 years and noone upgraded. Now in Windows Update Microsoft will force an upgrade silently in the next few months to IE9/10. http://windowsteamblog.com/ie/b/ie/archive/2011/12/15/ie-to-...
What I actually fear to become the new IE6 is Firefox 3.6 (and 4, to a degree). I've seen and talked to an unholy amount of people who flat out refuse to upgrade to the newest stable release, either because they assume (or have been told) that their add-ons will no longer work[1], or because they - for whatever reason - think the new release modell is bullshit[2]. I've rarely come across a reason that really wasn't just either (sometimes willing) disinformation or flat out bullshit.
[1] Which is stupid considering that the overwhelming majority of add-ons need exactly one fix, and that's the version check, which can be disabled with the Compatibility Reporter add-on.
[2] And then proceed to switch to Chrome. Oh the irony.
...which can be disabled with the Compatibility Reporter add-on
This is the issue. Most people don't know this. I was (1) an avid firefox user from the alpha stage and even I didn't know about the compatibility reporter with out some serious Googling, they should advertise it better
Especially since plugins are far and away firefox's strong suit over other browsers and they want to go with the new update scheme and change the version numbers so much.
(1) Sadly no longer am, due to the memory issues the Chrome team figured out how to fix. I still use it occasionally because I miss its much better user experience but until they get the multi process model done Chrome just uses resouces so much better I can't justify it
Exactly the reason that the latest Firefox now treats add-ons as compatible by default, rather than assuming incompatibility with each new version unless explicitly marked compatible.
That said, refusing to upgrade from 3.6 or 4 still seems crazy.
FF 3.6 is at 4% and dropping. People simply won't have a choice but to upgrade once market share gets low enough because developers will stop supporting it. Google will probably stop this year.
Switch to another browser if you don't like the new FF.
I'd assume you are talking about Google search? No they won't. 100% guarantee you Google will still support FF3.6, just like Google does, and always will always support IE6. jQuery, YUI, and all the JS libraries will always support IE6 as well. You should never go to any of these major sites and see JS errors & major rendering flaws in legacy browsers. If you do, that is just careless.
The reason I bring it up is because there is a biiiiiiiiig middle-ground of "support", and developers who don't realize this, need to. Legacy users don't need, and shouldn't expect the same experiences, but they still need support. You are still talking about Web browsers here. The Web doesn't have versions.
I know. If you are talking about some of the most advanced apps Google develops, I'm sure they will cease supporting FF10 & Chrome 20 at some point this year.
I prefaced my post by saying "I'd assume you are talking about Google search?"
The new release model is bullshit, since it carries no information.
But FF isn't so bad. Without the backing of big corp (which is what kept IE6 in this game so long) there is no or almost no pressure to make sure new homepages still works.
>The new release model is bullshit, since it carries no information.
You're talking about the versioning scheme, which I agree wasn't the best idea. However, the version number by itself never carries information. It's not an argument, it's nit-picking on trivial, completely unimportant details.
The release model itself (four channels, six weeks life cycle for release) is extremely good (there's a reason they copied it from Chrome, as much as I dislike the browser) and needed for a project such as Firefox. Longer release cycles just can't keep up with the development of the web in the long run, and provide much less direct feedback for the developers. Just calling it bullshit for no reason whatsoever is exactly the problem I was talking about.
For what it's worth, I suspect IE8 is going to stand out much more than IE7, which very few people will upgrade to (IE7 doesn't exist for Win7 I don't believe, so as people upgrade to Win7, you will see surge in IE8).
But nobody (we hope!) is writing IE specific apps anymore right? This is the whole reason IE6 is still around, because 10 years ago people WERE writing to IE. So I would suspect IE8 will have much less inertia.
I suspect it will be more about IE8's lack of support for CSS3 (and HTML5), meaning that we will have to keep bothering with various tricks and libraries to (partially) fix these for us.
I'm not trying to troll, but is IEN really just IE6? Has 10 caught up to Webkit/Opera/Firefox? One of the main reasons I feel that why IEX will be IE6 is its lack of support for webGL
They sponsored the last Hacker News Seattle event and made a big announcement about it then. Everyone was pretty happy to hear, and it was a great time.
Congrats Microsoft for spending so much effort phasing out an old product!
Yea, not only doesn't it support most of HTML5/CSS3, IE8 doesn't even support XHTML and also most of the DOM Level 2, both of which by now is more than ten years old!
3 ?
6 is dead 7&8 are still around and 9 is pretty decent. That leaves two more to go.
I wonder if ten years from now IE 9 will take the place of IE 6 takes today...
Nope, IE 8 will take the spot of where IE 6 is currently. It's the default browser of Windows 7 whereas one currently needs to update to IE 9.
The best hope to get rid of IE 8 is the same thing that killed IE 6 for good: popularity for the recent version of Windows. If Vista didn't falter the way it did, we would have had less time on IE 6, due to adoption of Vista's preinstalled IE 7.
Perhaps the solution is to develop a flash replacement for the HTML5 elements that we can then force feed IE and all you have to do is include that script on a page to give your users a modern expericence, even if they have to be dragged into it.
I love these sorts of maps, but I noticed something unexpected that I'd never noticed before: Japan's IE6 usage is almost six times higher than in the US.
For a country that I usually think of as "high-tech" (and the same place that gave us Ruby), that's not what I would've expected.
Is it true that IE6 has survived this long mostly due to a few 'go to' proprietary web based apps dependent on IE6 in the medical field? I've heard a few programmers repeat this. Curious if this is just misinformation being repeated.
This is the headline that so many of us fantasized about just a few years ago. The demise has been so slow that this barely feels like news. But with so many man-hours wasted on MSIE6 support, the headline's a legitimate cause for celebration.
If you didn't support any other browser, there by definition was no extra burden. In many enterprise shops you had IE and nothing else to worry about. I myself worked on apps where we only tested on IE because that's all the client used.
In fact this isn't true, because one thing that IE did better than any other browser was scriptable async http calls and XML transformation with XSLT in the browser. At the time no other browser could do that. We were using javascript to get XML data from the server using the ActiveX XMLHTTP object, and transforming it client side with XSLT to generate HTML. This enabled filtering, sorting, and different presentations of data all client-side. This was hugely productive compared to imperative-style Response.Write HTML building, and at the time was only possible in IE.
Seems like someone would look at that and consider America's advanced usage of the web as a potential competitive advantage, rather than rush through legislation to hobble it.
not to ruin it for you, but IE 6 usage is so high in China because China has a bunch of pirated XP SP0/SP1 installs; IE 7/8 is XPSP2+ only, and IE 9 is Vista+ only.
this same piracy is why the Business Software Alliance was originally a SOPA supporter.
It really depends on the type of customer you're targeting, but for most web apps and sites these days, I doubt someone still using IE6 is also a target customer for your service.
https://github.com/xdissent/ievms
(it recently added support for IE6)