IE6 gives me more stress than all my exams put together have given. I’ve given up on it because I’m absolutely fed up.

These days my CCs target IE in general but I haven’t really used them for IE6…

It’s good to raise the IE7 script again—indeed, I was so impressed by both Jeremy’s discussion here & Eric Meyer’s mention of it a few months ago that I decided to finally try it. The results weren’t great.

It’s understandable why, despite its promises, it’s not as widely used as one might expect. I wouldn’t recommend it to anyone; instead I’d recommend they start from Eric Meyer’s CSS reset & develop conditional stylesheets as needed.

The new design isn’t that bad in my opinion. Could be worse, could be better.I agree to point 1. Just try to make everyting work in ie6. As I already mentioned on 24ways, I’ll constantly check ie6 when building, and solve problems on the fly. If you first build the whole site and check it in ie6 afterwards it’ll take much more time to solve problems.

Oh my god!!
So “tenacious” IS a word!

An underlying factor to this whole argument about dropping IE6 support the fact that your average web user, or in fact most people do not know what a browser is, let alone what version of IE they are using. If the site displays incorrectly or looks wrong, they will just think it’s broken.

If your site is aimed at the web world crowd then fine, but on a larger scale, the internet to many is still just ‘the blue e’ on the desktop.

As you mentioned client contracts, here is what I say about the issue in my standard Stuff and Nonsense client contract. (Feel free to steal, change and reuse.)

“If the project includes XHTML or HTML markup and CSS templates, we will develop these using valid XHTML 1.0 Strict markup and CSS2.1 + 3 for styling.

We will test all our markup and CSS in current versions of all major browsers including those made by Apple, Microsoft, Mozilla and Opera. We will also test to ensure that pages will display visually in a ‘similar’, albeit not necessarily an identical way, in Microsoft Internet Explorer 6 for Windows as this browser is now past it’s sell-by date.

We will not test these templates in old or abandoned browsers, for example Microsoft Internet Explorer 5 or 5.5 for Windows or Mac, previous versions of Apple’s Safari, Mozilla Firefox or Opera unless otherwise specified. If you need visitors using these older browsers to see the same or similar visual design, we will charge you at the daily rate set out in our original estimate for testing and any necessary additional code.”

No matter which way you decide to go, the person hours spent working around flaws in IE6 (and NN4 before it, as you point out) is phenomenal. I prefer the write-to-standards-and-let-god-sort-it-out approach myself, but for some reason that doesn’t always go over well with the suits.

I was just talking about this for a friend of mine who’s interested in getting me to code for him. Quite a coincidence actually.

Anyway I love your equation that’s just brilliant. I’ve never tried out the JavaScript method before I usually use the conditional style sheets. I was thinking of actually using the script from IEDeathMarch on my own site. But I think instead giving ie6 users a more basic version of the site is a better option. How basic I haven’t figured out as yet.

Wow imagine having to whip up a design just for IE. Dam we’re really going to have to wait till forever to use CSS3 aren’t we…

Excellent bit of contract banter there, wing commander Malarkey. Permission to borrow/plunder/shamelessly steal?

Good discussion, good framing of the issue.

My first comment on the site regarding an excellent article which should be an important issue to us all.

The equation that you and Natalie devised is brilliant and something I’ll be brining up in work on Monday morning. I really can’t wait to read what others have to say on the subject and article as a whole.

Andy – thanks for offering up your contract text, it’s something which I’ll probably use myself from now on.

While conditional comments have gotten a lot of play, I’ve now switched to using them on the BODY tag to set a class.

This has two benefits:

1) One fewer http request (sometimes two if you have ie6.css and ie.css)
2) You get to keep your CSS in a single location and not split up similar rules in multiple files

I have more details on this condition comment body classing technique at paulirish.com

nice.

one addendum i might add for a minor situation: i have used the google script, and it works well until you’re in a secure situation.

If the script is going to be used in a https environment, IE throws an alert about the document containing both secure and non-secure items, which might scare off some buyers (a reason you might be secure in the first place).

it is easily worked around, if that’s an issue, by putting that script locally on your server, and then accessing it with a relative URI/URL (/js/script_name.js). At that point, IE won’t throw the alert.

I agree with the previous comment. That’s a very practical approach. Also, I had not heard of the IE7 Script and I’m thinking why not use that and see how it works for most of my projects. I don’t want to spend a lot of time debugging for IE6 or learning the art and science of same. I crave the script.

I like to think I’m good at math, but in the time it would take for me to figure out that equation, I could have IE6 hacked up and ready to play ball. :) This is a good read, and I wish I had invoked Yahoo’s ‘graded support’ approach to my current work months ago. I will definitely do so for any future project.

@malarkey – well-said, sir!

I totally agree, I think that a recognised set of browser support standards would be very useful.

As you pointed out, Yahoo set the pace or at least popularised the idea of graded browser support (gbs), I think the BBC’s Browser Support Standards Work Group have taken it further with clearer definitions of those levels of support:

Browser Support Standards v3.3

I think having a common set of support level definitions would help manage client/stake-holder expectations.

Andy Clarke has written some excellent posts recently on browser support and related issues with client/stakeholder expectations, these being very good:

Time to stop showing clients static design visuals

Five CSS design browser differences I can live with

Maybe along with ACID3, a standard support definition could be hosted on Wasp.

Nice post Jeremy.

brilliant article Jeremy. Clear & consistent communication pays off for everybody, ie6’s days are numbered.

I agree that for now at least, the answer lies in some sort of moderate level of IE6 support. As much as I would LIKE to take drastic measures to promote upgrading, implementing a redirect to download the latest version of IE is not the best idea…

While I reluctantly “support” IE6, I’ve definitely toned down the time spent being anal. The site doesn’t have to look identical anymore, just close enough – unless the client wants to pay more.

As it stands right now though, “support” should mean usable, not pixel perfect.

That said, I will take my liberties using CSS 3 properties when it makes sense, because this makes the site better for peeps using modern browsers. (Sorry IE6 and 7 users, but the border-radius property beats cutting up images for rounded corners or using javascript hands down. Guess you’ll have to see 90 degree angles…)

BTW, I think the formula is awesome! I might find myself plugging numbers to see what makes sense.

Of course, if the day comes when we’re all going to band together in a civil disobedience and drop support completely, please someone make sure i hear the news :)

Excellent. I’ve often battled with the option to use the IE7 script, having yet to give in to the urge. Your equations are spot on, and I will definitely use them for my next project. Along with @malarkey’s contract lingo… brilliant.

I couldn’t help but notice in the argument two tiny little things were ignored all together. UX and Accessibility. At what point do file size / download time / avg user connection speed and JS disabled rate come into play.

Hate to nitpick, but the aforementioned items could play a major role in the appropriateness of Dean’s script. IMO.

Great article on a topic that has been plaguing my conscience as of late. I actually just ran across Dean Edwards’ javascript method last week, but have yet to try it out. A quick question for any of you that have had experience with this method:

When this method effectively up-versions IE6 to IE7, if IE7 is not minding its manners to begin with (while being worlds better than IE6, IE7 is still rather imperfect itself), do you still have to make an ie6.css to fix the IE6 inconsistencies, or is there a way to get IE6 to read and implement your ie7.css fixes as well?

>>>Block IE6 users from your site
Block 20-30% users from your site??? Good idea)) It’s a joke, isn’t it?

I always use the “CC” method to include extra IE stylesheets. And sure sometimes (or should i say: always?) i damn the exist of IE6 but shouldn’t we know the special hacks/workarounds for IE6 by now, after so many years developing?
Give IE “HasLayout” and you can fix 90% of misbehaviour. Sure you need more time to make websites look nearly the same in IE6, but our clients are willing to pay for it, because 20 – 50% are still using IE6! That’s worse, but true…

@Paul Burgess – agreed to a certain extent, but those are the same people who, to generalise, will already be on IE7 because they always click “okay” when offered updates etc.

I’ve never used the IE7 script, but only because I develop a slight tic when using JavaScript to help control layout and design. IE6 persists most (I’m guessing) in corporate environments where updates are managed centrally. These are the same environments where we are most likely to also encounter disabled JavaScript, and so using JS to workaround IE6 deficiencies simply adds to your progressive enhancement bottom line, as you still need to include an IE6 “only” workaround stylesheet.

As for the rest, I think this is a great discussion. Finally, it’s something on which most web developers seem to agree, at least in principle, which makes a refreshing change!

Wish there was some way to subscribe to the comments…

As per the previous comment, it’s worth taking into account the possibility that many IE6 users are corporate users who are locked into their browser through no fault of their own. These users may also have a higher than normal chance of being behind a firewall that restricts or removes JavaScript, rendering Dean’s script solution useless.

The equation is perhaps not as simple as it seems.

I like the Malarkey approach of charging more for ie6 support, the equation explains clearly the poor ROI of supporting the outdated browser.

If more sites receive degraded support in ie6, then people, and managers, will be more onboard with flushing it out of the market. I mean it’s not economic exclusion here — browsers are free to upload.

I wish IE6’s days were numbered – I just checked the stats for a website with a worldwide and mainly corporate-only audience, and nearly 50% of the visitors are using IE6. Looks to me like corporations are very reluctant to upgrade their infrastructure. Which is one of Microsoft’s big problems. And for a website with a UK home audience, just over 25% are using IE6, down from 33% at the beginning of the year.

I see myself designing so IE6 looks and performs very similarly to IE7 for some time to come. I don’t see a time versus effort tradeoff usually being practical for IE6 just yet. Same goes for IE7 (which also has its frailties). Hopefully, IE 8 when out of beta will perform as well as the other web standards compliant browsers (which is what I state that I design for).

My experience with clients is the same as Paul Burgess’: they don’t know about the problems. And they aren’t really interested – they just want a website which reflects well on their business. So, along with as much education as the client will take, I’ve been trying to limit my risks, in writing, without making it into an issue.

Seeing what Malarkey uses contractually, I’m encouraged to be a lot more specific. Browser bug fixing can be a significant cost and time problem. I also include this line (I’m referring to web standards compliant browsers) “The website should be forward compatible, but as we have no control over browser makers, we cannot guarantee this.”

Great article Jeremy.

Some good points there, I like the ‘tough love’ approach on not spending hour after hour trying to fine-tune IE6 experiences to match modern browsers. I guess that sends the wrong message to people still using IE6 that it’s OK to use that browser.

Excellent article, Jeremy: everything you write is always so clear.

Nevertheless, I think Antoine has a point (accessibility). CSS and JS are different technologies with different support and availability and it’s not always a good solution to use one of them to solve the problems of the other.

Somehow, the formula should include, not only the portion of your audience using IE6, but also the portion using JS.

Paul Burgess said “An underlying factor to this whole argument about dropping IE6 support the fact that your average web user, or in fact most people do not know what a browser is, let alone what version of IE they are using. If the site displays incorrectly or looks wrong, they will just think it’s broken.”

That’s equating “dropping IE6 support” with “the site displays incorrectly or looks wrong.” Is that necessarily the case?

If we’re talking about what Dan Cederholm calls “progressive enrichment” then what the IE6 users will see is a slightly duller version of the site with fewer text effects, fewer rounded corners and fewer gradients. Does that mean that the site is displaying incorrectly?

If, on the other hand, we’re talking about layout issues where content overlaps or is clearly out of place then yes, people using IE6 will think the site is broken. But what is the solution in that case? To just fix those glaring problems or to try to make the site look exactly the same in IE6 as in all other browsers?

I’m throwing these questions out there to point out how unproductive it is to talk purely in terms of “supporting” or “dropping support”. We need a wider vocabulary of terms than that.

I like Lauren’s short and snappy definition that “‘support’ should mean usable, not pixel perfect.”

Antoine asked about the user experience for IE6 users and how that will be affected by using the Dean Edwards script. Well, that’s where the equation comes in. The download time for the script is a penalty that IE6 users will pay. To figure out whether or not the penalty is justified, plug your numbers into the equation. If you get a low result, it isn’t worth penalising IE6 users with the longer download time; you’d be better off spending some time on IE6-specific development.

I’d been toying with the idea for a while now wether to drop IE6 support or at least charge more for making it look as the designs intended. I opted for the charging extra idea because as bad a browser as it is; clients need the sites you design for them to be visible to as many people as possible and all the time IE6 still commands a reasonably high user base then IE6 will continue to survive and be a pain in the proverbial.

The time it takes to bug fix things in IE6 is a real resource hog and at least this way i get to charge for my time as opposed to making a loss on a project.

I think that the structure(grid) IE6 should remain stable. I don’t care about small differences. But if the structure brakes the site looks not “professional” and you will lose all 20 to 50% of IE6 users.

You should read our blog post about why we continue to put up with IE6: When can we stop pandering to IE6?

Although I’ve never used the IE7 script before, one other thing that may be worth considering is that even if you use conditional comments for IE6, you’re still not going to be able to have access to all the extra CSS2 goodness (e.g. attribute selectors) without javascript. So if you like to use that in your projects on a regular basis, it may be another tipping point in favour of js over conditional comments.

This is a pretty vital discussion, and one I rub up against every day. The answer has to be something that can be easily understood by everybody involved: clients, account managers, back-end people. Just today a colleague was confronted by the requirement to sort out a strange peekaboo background image error in IE6. It seemed pointless to be spending effort “fixing” such a thing – the only real fix would have been to remove the background image altogether. The point is: this lack of a common understanding is a waste of everyone’s time. We all should “support” IE6 – our clients’ sites should be usable, and should take money, on that browser. But we should also promote maxims that blow apart commonly held misconceptions. Most people don’t hold it as a truism that web pages don’t have to look the same in every browser – to many people this is, by default, what “support” means. So step one should be scoring this on the minds of all around us:

Web Pages Do Not Have To Look The Same In Every Browser.

…then we can go from there.

My approach to IE6 is now along the lines of this:

1. Layout should be 100% correct
2. Typography & images should be acceptable and not ugly (e.g. fixing borders, decent gifs to replace transparent pngs)

So basically what Dan C. says via Jeremy above—nothing’s broken, but nothing has to look too fancy either. The site has to work and not be ugly.

I’ve taken a look at IE6 use on my clients’ sites thru Google Analytics stats, which reinforce my gut feeling— visitor use of IE6 is very dependent on the target audience. Sites I’ve built that have a primarily young and/or artistically-inclined audience have virtually zero IE6 use. Sites visited by an older, more mainstream audience have 20% – 30% IE6 use.

So as with any other aspect of your design strategy, it’s essential to know the target market for the site.

Nice article! To be frank, at first, somewhere in the middle of the article, I was amazed that a complex mathematical formula was introduced (which I hardly understood).

Then, towards the end, things were simplified and basically ended in “Be wise, use your own judgment, when dealing with IE6. If your audience is ~30%+ of IE6 users, and you have a lot of time for developing the website, use conditional comments for IE6 and write your own CSS fixes for this old browser. If, on the other hand, the IE6 users are fairly low in number, and/or you don’t have too much time for development, then use the IE6 script.”

For my part, I can say that I am almost as eager to see this old browser go away, as I was eager to see Netscape 4.7.x fade in history… :-)

Nice approach using that formula! Made me laugh a bit in a positive way!

I mostly check my sites in ie6 while building them and not after I’m finished. Taking time to fix things for ie6 depends on how important I think it is that they’ll work properly in ie6. But I never spend too much time on it.

We just keep on waiting for the day we can bury IE6!

Why are so many people still using IE6? Simply because the sites they go to work and there is no reason for them to upgrade.

For the average person, they look for the IE symbol. That’s their gateway to the Internet. They have absolutely no idea that their browser is out of date, nor do they care.

Most people only change their browser when they change their operating system. Which is when they buy a new machine. That’s probably every 4-5 years. By 2012 IE6 will be where IE5 is right now. Which won’t be any good for developers if the Mayan calendar has been interpreted correctly, lol :(

If we want to get people off IE 6 surely prompting them to upgrade in some way is the way to make this transition happen sooner rather than later.

Have you seen that excuse for a browser that is IE Mobile (or whatever it calls itself)? It’s awful.

I think another part to this that has not been mentioned is educating clients and bosses, as well as the general user, that ‘Progressive Enhancement’ so to speak is the way forward. In large corporations and across the clients I have worked with there is still the mentality that all sites have to look perfectly the same across all browsers.

We as Web Designers have a responsibility to push this education out to people otherwise we are just making our job harder. As Malarkey said above even adding this into the contract can help to bring these issues out in the open.

I have been working on a Beta site recently, that had a very strict time frame that would have totally been blown out of the water if we had worried about IE6 whilst designing and developing the site. Due to this I convinced the bosses to send all IE6 traffic to the old site (bear in mind we still have over a 30% user base in IE6), should this be an issue? No, the users can still access exactly the same content but just not with the funky AJAX content and advanced layouts. Once we go out of Beta we will push users to upgrade their browser to view the new site or be stuck with a downgraded version. The choice is up to them.

To be honest, I’ll be sad to see ie6 go. No other browser has, or probably ever will teach me more about css than it has.

We only understand how anything works when it falls apart and we have to put it together again.

Ok, I have attempted to run the calculation for a recent site (http://bookfairs.scholastic.co.uk)…

80 total dev time
10 ie6 dev time
13399 total audience based on 30 days usage
5331 ie6 audience

p=24.859

What I don’t know from the article is — what does it mean. ~40% of our users are on IE6 (mainly teachers in schools)? Does that number suggest we should be using the script or not?

Thanks for any help,

Peter

p.s. prompted by a Malarkey tweet…

The answer to “support” with my clients has been to produce an “identical replica” of what works in other modern browsers. This does restrict what is possible with the designs to a small circle, but nobody is willing to consider degraded support for IE 6 claiming there are a lot of people still using IE 6. I wish Microsoft would be forceful about upgrading browsers to IE 7 when they are selling their new OSes as well.

Personally I think we’re still years away of being able to drop support for IE6. Its market share is still substantial on all websites I manage. The percentages are dropping, sure, but at a very slow pace.

That said, I don’t see the problem with IE6. I find its layout issues to be fairly predictable and to mostly have simple workarounds.

Compare the current situation where only IE6 is a bit of a pain to about 5 years ago where we had to support IE5.x and Netscape. Now those were really painful times.

Simon.

I agree with http://www.catchmyfame.com/2008/12/02/attention-designers-stop-supporting-ie6-now/. Stop pandering to IE6 in any way and it will speed its demise.

Woah, the good feeling to see IE6 being put out of its misery.

It’s a shame though, a lot of places I’ve been to, mostly places where many people use the computer still use IE6. For example, my school still uses IE6, and numerous Internet cafes I’ve visited have it.

I asked one of the technicians why we still have IE6, it’s because the upgrade process is too complicated considering the additional security software used in a school environment.

Nothing I fear more when testing, the IE6 run. Nice to hear about the script that emulates solid CSS support, might give it a go next time round.

@Jonathan Barrett That’s the common wisdom that a majority of IE6 users are coming from corporate users with controlled updates. Be interested to know if there are any numbers out there on this.

—

Equation but is cute but I think that noting that browser support isn’t an all or nothing proposition is valuable. Personally I find myself styling for IE6 just so nothing breaks.

—

Thanks @Malarkey and @Paul Irish – found your comments interesting.

Sorry to join so late…

I have a background in software training and it turns out that naming your courses is very important. If you have three courses pitched at different levels you can’t call your intro course beginners or basic because people tend to think they’re already beyond that and sign up for the intermediate course. It’s then a real pain in the neck when you have to go back and teach the basics. So we tend to call the course “An Introduction to…” or “A Foundation Course in..”. The next two levels enforce this by being pitched as “intermediate” and “advanced”.

Taking a very simple overview of browsers as things stand we also have three levels: Basic CSS (IE 6); Good CSS (IE7); Great CSS (Firefox, Safari, Opera, IE8…). For these three levels I propose these terms

Foundational Support
Standard Support
Advanced Support

I’m not 100 percent happy with the term Standard, but other options like Enhanced or Enriched are already in use and it wouldn’t be obvious which was the highest level.

Foundational sounds much better than Basic,who wants a basic web site! It also says that we will be building a strong, unbreakable web site.

Getting IE6 to behave is always a triumph or a relief,,, especially when using floats or positioning.
Now hoping more advanced CSS selectors find their way in IE7 and IE8 as well.

I still develop and test for IE6, I don’t really have much choice like most people still but I think my new year resolution may be to start phasing it out and charging extra for exact layout matching.

One of the things worth noting is that most new computers come with Vista these days and IE7 is the default on that so it’s being phased out in that way as well as whatever other measures people take when building sites to “educate” visitors there are alternatives.

I’ll throw in another possible option which I mangled together using a stylesheet switcher from ALA and conditional CSS and that allows me to show an upgrade notice but allow the visitor to dismiss it if they can’t or don’t want to upgrade from IE6. You can find the instructions and relevant files zip on my site here if you want to check it out: http://www.welcomebrand.co.uk/12/11/2008/encouraging-a-soft-upgrade-from-ie6/

Also, thanks to Malarkey for the contract snippet, I’m definitely going to be using that in future!

J.

Great article! I know how I’m gonna do with my site which I’m going to build. However, since it’s my personal site, it will not always be possible to do like this when your working for a client, or in a web development company.

1. I will present a splash page for IE6 users (perhaps IE7 as well) which shows up at the first time a visitor accesses the site saying something like this:

Your web browser, Internet Explorer 6, does not follow the XHTML standard, which is used to display web pages properly. This means that the site will look strange. In order to give you the most pleasant layout, please consider using another web browser, i.e. Firefox, not only to enhance the experience on this site, but on all sites. However, you may proceed if you want to, but the site will look fragmented and strange. [Link to mozilla.org] [Link to proceed]

2. I will have a red cross at the bottom of the page. Next to it it will say something like:

[x] Your browser is not fully compatible with [site name].

You can also click on it to display something similar to the splash screen.

Thanx

Very Interessting Article

Truth is IE6 user base is at 18.5% according to w3schools. That is about a 40% drop from last year in same month. Even though it is a sharp decline I feel that 18.5% is worth keeping the IE6 workarounds still in place. We all know what to expect with IE6 and we’ve all had to deal with the issues. I don’t believe there is that much more effort required to make that % unworthy of our time.

One thing that is left out of this article is the demographic influence. If your site/store targets people who would be in that IE6 user base…it would be well worth your time to support it. If you know your demographic would have nothing to do with IE6 (such as this website’s target demo) then great. No need to worry about IE6.

For me, until I can look at the IE6 user base and justify losing those customers, I’ll keep dealing with the headaches of IE6.

Just my opinion.

My opinion is simple. Microsoft is the most profitable company in the world, they can pay large salaries to their emplyees.

My responsibility is to build XHTML that follows WC3 standards (don’t try to verify my site, I’m still developing a new theme to it and then I’m gonna fix plugins and update posts :P ), and theirs is to develop browsers that follows these same standards.

If they, being so well paid as they are (should be), don’t follow the guidelines… why whould support their inconsistences for free?

I’ve also read about another way to drop support, simply don’t provide any CSS to IE and let its faithful users browse raw HTML. I just don’t bother with it, let they see how their beloved browser deals with open standards. If they want good support, let them donate enough amount of money so that we feel worthy supporting.

Last weeks I went to a site that PHP-verified I’m using FireFox and just sent me an HTML saying my browser is not supported…

For a site like this a lot of visitors would have firefox, designers have moved away from IE. But for sites like facebook, myspace, where people are totally oblivious to this it is important to include IE users. Current they make around 33% of web traffic (dec 09) with 10.5 % using IE6.