With IE8 promising full CSS 2.1 support, and the beta version already passing Acid 2, it seems that we are finally coming to the end of hacking for IE. But how long will it be before IE6 dies? What about IE7?

It's now three years since I first taught myself standards-based, "table-less" web design. Firefox 0.8 was released more than four years ago, in February 2004, which we can take as an arbitrary date for the beginning of the standards revolution. Throughout this time, Internet Explorer has aggravated me -- and you too, I imagine. I've had to waste my time hacking for it, learning arcane tricks that have no useful application beyond pandering to a badly-behaved, backwards browser.

Thankfully I no longer need to hack for IE5; it's dead and buried. But IE6 retains almost 40% market share.

I'd like to ask some questions of you:

1. Do you think IE8 will achieve a level of standards fidelity on par with current browsers such as Firefox, Opera, and Safari?
2. How low must the share of IE6 drop before you stop "supporting" it?
3. When do you think this will happen?
4. What about IE7?

And here are my current thoughts on these questions:


1. IE8 standards fidelity

Yes, I think IE8 will be the first version that has complete (or near enough) support for CSS 2.1. It's not clear which browser will have the best standards support, but I think IE8 will be the first version of IE for which I do not need to hack.


2. IE6 share

I think I will stop supporting IE6 completely once its market share drops below about 5%. It's just not worth the hassle to me.

However, that decision will be influenced by individual site stats too. Also, I'll begin dropping support for IE6 earlier -- say, when it reaches about 10%. I'll still hack for it, but more quickly and with less care for the end result.


3. When will this happen?

That's the big question, of course.

I suspect that a large proportion of IE6 users are from business PC's. Large businesses do not update their end-user software often. Instead, they maintain a standard image for all their desktop PC's. This is a complete hard-disk image, with all the software, drivers, and whatnot. When adding a new PC, they simply download this image.

(By "large business", I mean "company in the top 500 globally for turnover". And by the way, Microsoft is completely dominant in large-business end-user operating systems, mainly due to MS Office.)

This policy leaves large companies far behind home users in adopting new software. Many large companies are still running Windows 2000, for example. The advantage is that, by standardising their software and drivers, they greatly reduce the complexity of technical support. When you have 120,000 PC's to support, these things matter.

A typical image lifetime is about four--five years. The development of a new image will take about a year (including testing and deployment). Only software that has been stable for about six months will be considered for inclusion in a new image.

(In case you're wondering how I know all this crap---blame my dad. He's arguably the world expert in end-user services.)

So let's do some calculations for the expected upgrade of business users to IE7. IE7 was released in October 2006; add six months to that and you get April 2007; this is the first date from which businesses would consider including IE7 in their next image.

In April 2007, some businesses will be ready to start developing new images, but (at the other extreme) others will have just completed a new image. By October 2009, about half of businesses should have included IE7 in their new images; and by April 2012, all businesses should have included IE7 in their image. Add on one year for image development and deployment.

So, assuming business use is (almost) entirely to blame, we would expect to see IE6 declining from April 2008, and finally disappearing by about April 2013. Assuming a linear decline, IE6 would lose about 0.63% market share per month (taking TheCounter.com stat of 38% for March). This is consistent with current trends (although that doesn't prove anything).

If we stay with my model, we can predict that IE6 will drop to 10% market share by December 2011, and drop to 5% share by August 2012.

So I predict it will be over 4 years before I can drop support for IE6.


4. What about the death of IE7?

This is harder to predict, because we don't know what the stats will be when IE8 is released, and we also need to account for all the home users. Nevertheless, we will not be able to drop support for IE7 until most large businesses have upgraded to IE8.

Let's assume that IE8 will be released in November 2008. Add six months for stability: May 2009. It will take about 5 more years for business users to upgrade, so we can expect to witness the final gasps of IE7 in about May 2014.

Depressing, isn't it?

Hmm, some interesting points. Here are my thoughts.

1. IE8 standards fidelity

The fact that IE8 has passed Acid 2 at this stage in the Beta process is a very promising sign and I too believe that IE8's much trumpeted support for the standards and bold claims such as "IE will be the most standards compliant browser to date" will put an end to hacking for IE. Of course, the end to hacking will only really come when the legacy browsers truly die.

Certainly, having all the major browser players supporting the standards of web content rendering will make it easier for us as designers to put together websites but I fear that the chasing of standards has become the only goal being aimed for. It goes without saying that a browser should support the W3's recommendations but if you spend 100% of your time chasing the standards, conformation rather than innovation becomes the new habit. Imagine if in the beginning there were no browser wars, imagine that IE and Netscape didn't battle furiously to gain the market share in the first 10 years of mainstream internet and imagine that right from the outset both IE and Netscape met the standards (of course there weren't any at the time but if there had been) with the only change ever coming from the and update to the W3's specification. Innovation in web content may not have been as rapid and modern day staples like Javascript (created by Netscape to encourage web authors to gear their sites towards Netscape) and Flash (developed in part to allow advertising and content to display independent of browser and OS) might not have come to be.

My point is not to buck against the trend of the web developer community and it is certainly not having a shot at the W3, it is to say that I think standards compliance has become too much a be-all-end-all mindset and less a solid framework from which to launch innovation. As builders of websites we cannot really create any revelations in the content we display, we can use things like Javascript, flash and ajax in new and exciting ways but we cannot implement a new concept in markup structure or document rendering, this has to come from the browsers. Because the browsers chase the standards and stop when they reach them, we also must conform to them. I just think that a little bit of intellectual sparring and market-share-chasing innovation with the testing of proprietry code is a good thing provided the groundwork is still there by meeting W3 standards.

2. IE6 share & 3. When will this happen?

I agree with your points here completely. My last occupation was within a company that only last year upgraded employee browsers to IE6, well over a year after IE7 had been released. This just made me scratch my head. It probably went hand in hand with the decision to keep users running Microsoft Word '97 until the present day.

The day IE6 dies will be a happy one for the web as a whole. There is a way for this to happen sooner rather than later and it could be that we just stop coding for it. True, IE6 may have a 40% market share but of the total amount of internet users, what percentage still run IE6 and use a computer that will not support IE7 or FF? I would be willing to say they would account for less than 10% of the total internet populace. Browsers are free, the internet is obviously available to these people (or else why would they need a browser anyway?) and the least favoured of the two options (IE7) comes both in an offer from Microsoft update and a reputable website that even the biggest fraud-a-phobe could feel moderately comfortable downloading from. Agreed, many computer users are "happy with the way it came" and don't proactively seek to change their software but I feel that coding for bad browsers is the same as leaving newspaper around the house for an untrained puppy; you could keep leaving newspaper on the floor so when the puppy made a mess it could easiliy be cleaned or, you could train the puppy to go outside and never have to put paper down or clean up again.

I am not far away from training the puppy when it comes to IE6 and in a couple of months will probably stop supporting it for all but the sites that will target the larger companies with their mass-deployed computer images.

4. What about the death of IE7?

IE7 has not been so bad for me at all. Obvioulsy it has quirks and glitches like the older ones but when running in compliance mode I have found that it generally holds up okay. The hacks and changes to get IE7 working aren't nearly as time consuming as they are for IE6. I feel the death of IE7 when IE8 is realeased will come, comparatively, a lot faster than than the death of IE6 after IE7. The reason is that a lot of people are still using IE6. Why would you upgrade from one legacy browser to another?

When finally, those people chugging away with IE6 upgrade they will jump straight to the standards compliant IE8. For a lot of people this will come in the form of a new computer or a Microsoft update but I do believe we won't be wating anywhere near as long for the grass to start growing on IE7's grave.






I am happy to see IE8 on the way because it means browsers are on relatively equal footing again and nothing sparks innovation like stiff competition.

I agree with all of the above two posts, but in my opinion i don't think you can generalise too much. 40% is, admittedly, a lot, but i think that you also need to look at your own user stats. Having software such as analytics is very useful for indicating how much hacking you need to implement to get your site IE6 complient.

Some site will have a very low amount of visitors, sites for web designers who know the advantages of using firefox will have a much less percentage of IE6 users than a "run-of-the-mill" forum or information site.

My thoughts on the matter lol

Agreed. My stats give me about 28% IE6 share, which unfortunately is still far too high to ignore.



A very interesting point. Wars advance technology, and the browser wars were no different. I'd rather live in peacetime, of course!

We need some arena of combat for browser vendors -- some space in which they can show off cool new proprietary technologies, and gain kudos from it. And then we need a way for those technologies to be integrated into standards.



The problem is that you can't make that assumption. My badminton site gets the fewest visits during the weekend. I can only explain that by supposing that people are spending more time browsing it from work than from home!

Your site may be irrelevant to business, but people are still goofing off work to browse it. But as Craig said, you can use your own stats as a guide here.

Do you mean this literally or in context to browsers lol?

In relation to the browser ways, there are two sides. On one hand having to code your site to suite so many different browsers can be time consuming and at times very frustrating. But, imagine a world with only IE, if you think about it you would be viewing the web through the eyes of microsoft, a good or bad thing...

Both.

The browser wars may have stimulated web technologies, but I'm glad the wars are over. I'm much happier making websites in a dull, standards-compliant, peacetime world.

Hopefully things will be a lot easier in the future

If you try to update XP then you will be prompted to update IE6 to IE7. Vista ships with IE7 by default. I would hope/expect IE6 to be near irrelevant within the next year.

It makes it a touch harder when placed into the real world context. I would much prefer that all the wars were over in life, even if they do in fact advance technology. I guess my point is that I want a war that has no effect on people who don't want to get involved. If both browsers meet the standards and then go to war on some big new thing, the web developers who stick to the baseline and code for the specs won't be affected. If a web developer chose to jump into the war ad start using the new browser weapons then they cannot complain if the new technology dies.

I think is should be possible for new technology to be placed in bowsers without ever effecting the baseline standards. Put simply, if users like the new feature that browser's competition is bound to take notice.

I know, I too am guilty of sneaking onto the internet from these production line software setups. What's to say that dropping support for IE6 won't dramatically increase productivity in big business? No longer getting eye candy while slacking off so better get back to work and wait until I get home to do my personal browsing.

Having had some experience with large scale corporate systems, I know that it is possible to deliver a software upgrade across the netwrok and that it is not difficult to set this up but as you say, the whole point of the images is to reduce tech support so I guess there's no one that can do this in a lot of cases.

I hope you're right, but I expect you're not. Businesses simply don't upgrade that often.




Yes, the reason for images is to create a stable, safe desktop environment where everything has been thoroughly tested. The ideal is to have completely standard hardware and software. Tech support is hugely expensive, so it's just not worth messing with this stable setup.




Yes, and browsers have actually been doing this already to some extent. Take -moz opacity, for example. Or IE8's webslices.

Consider also that CSS 3 is not yet a recommendation, it's a working draft. Yet browsers are already implementing elements of CSS 3. Even CSS 2.1 is only a candidate recommendation!

I thought it might be interesting to look at the trends. Here's a graph I made, using theCounter.com data (see attached).

You can see that overall IE share has remained largely unchanged, albeit with a strange dip between March and November 2007 (the drop is from IE7; perhaps this is due to home users uninstalling it).

IE7 overtook IE6 in February 2008, but since then the graph has flattened off and the gap has widened by only two percentage points (meaning only one percentage point decrease in IE6).

That's a disturbing plateau for IE6 from February 08 to now :'(

Great posts. So in depth!

I chime in with the idea of training. We started plugging firefox quite offensively a few months ago. If you have ie6 on our site in different places you will be recommended that a better experience is available. We say "ie7 is available, or get firefox".

We say "sorry, this site doesn't look correct for you because you're using an insecure and inferior browser". In slightly less harsh words. But its the truth. And why beat around the bush? Thats what bush did.

What we developers must realise in my opinion is that the longer you support ie6 the longer you hold it all back and the longer you actually spend time training yourself into a formula that is disfunctional. Sure, business might be bleak for a while because half your ie users get distracted in a wonderland of firefox or ie7, but you know what... they will remember you the next day when they load up their awesome browser again "thats right!!! my life is awesome now because of firefox... who told me about this? I'm so glad that person pushed me to make the shift... wow... what do they sell I can see already through this free upgrade they know what is good".

Supporting ie6 without a notice of a better browser (in my opinion) is like getting a child into your care, and not speaking ever. There's a whole world of sound to be experienced... but you don't teach them that. Just let them grow up in silence. Just doing hand signs and drawing on paper. In that situation, it doesn't seem right does it? But really, its not that different... your web visitors are in your care also and I think what appears to be a fear of making a stand in your business or to your clients about moving the paradigm beyond ie6, then you're not bringing the revolution any time sooner.

I hear the buddhist believe that if you are aware of something somebody should know, that its not good karma to not say something. If its something you feel strong about that causes you disharmony its up to you to say something. To make a stir. So to just sit there plugging away this old paradigm and not saying anything I think is a real disservice to yourself, and your clients.

Consider 'browsers' aren't taught in schools. Who are the internet teachers today? We all are. I really think the movement from ie6 will happen faster than you think. Consider 100th monkey phenomena. Critical mass. Flowers blossom in what seems an instant relatively to the time it takes for it to grow..

This post has been edited by cosmicbdog: Dec 10 2008, 10:01 AM

Having a message about IE6 on personal sites is one thing, Having it on business sites is another... why would you make customers have either a poor user experience or navigate away from the webpage... it just doesn't make sense. You are more likely to damage your business than empower users.

I think that by all means there should be features which are not available to IE6. Not sure what exactly but a message saying 'IE6 is not compatible with this section of the website. Please upgrade to internet explorer 7.' would be acceptable. The problem arises when large sections of your site either display incorrectly or are inaccessible. That is 30% of the market you are missing.

I am a firm believer that people should have the choice to use what they want to use. If people want to use IE6 so be it.. it will just cost site owners a little bit more money to make their site compatible.

Take software apps, software is made for current Operating systems. It would be silly to make Vista only software when a large proportion of users still have XP. Furthermore aren't IE6 the most valuable users, they are about the only people who won't be charging around the web with an ad blocker and other silliness.

I posted on your blog about this as well..

Why is it silly to let people upgrade their browser? It's free, isn't it? When there's a new version of Flash: you let them download it, new version of Java: download it, new version of your anti virus: update it.. But a browser is sacred and should not be touched? People get requests to update/upgrade their software all the time, but nowadays the most important one (after anti virus), the browser should not be upgraded?

When you don't update your anti virus software: you get viruses. People take a risk not updating. People take a risk of not being able to view a flash website when the don't upgrade their Flash version, but a normal website that looks a bit wonky here and there (because if you code properly, it's not like the whole website messes up, it's just minor things like non-transparent .png's), and therefore still accessible should not be upgraded?

Your comparison to Windows doesn't make sense.. I can't run software made for Windows 95 (or Windows 98 in mos