Should I support all (6, 7, 8) IE versions?

Question

I have FireFox, Opera and Chromium web browsers on my Ubuntu box and I manage to run cycles of development and testing pretty quickly (and happily).

The best part is the small CSS files I mange to produce for all types of layouts.

The problems always start when I decide to check how I am I doing with IE, and I have a few boxes just for that matter (got tired from all those VM games)

Anyways, trying to support those nice layouts on IE take a huge amount of development time, make my CSS files ugly and even force me to compromise on the end design.

I've tried using some libraries but in the end they all leak as other abstractions, leaving you the hard debug issues anyways.

I'm not sure I can afford this but I consider avoiding IE at all ! and yes I know the statistics and yes I know that your clients will kill you but I'm sick of it already!

What if the entire development community stop supporting IE with all its bizarre versions? maybe its time to convince those guys at Microsoft to concentrate on the single thing they ever did well (which Joel remember so well) and leave us alone already?

Go ahead and crucify me with "This is not a question" etc. I'm just trying to legitimate my decision.

Guy

Answer 1

The real question is what is your audience. If it is an intranet app you have some control over what browsers they will use.

If your app is in the wild, then it is a question of how much of your audience you can afford to give up on

Answer 2

I suggest you design some conventions, that will cause your HTML/CSS be both standard-compliant and render well on IE, and then follow those conventions.

From project management point of view, you might first make the core version for standard browsers and mark that as a distinct achievement, and then proceed to improve for marginal browsers.

Answer 3

You're not the only one who feels this way. I would suggest looking at the traffic for the sites you create--the browser breakdown can vary wildly from site to site. If you don't get much traffic from IE, then go ahead and stop support!

Answer 4

IE sucks, we all hate it, but the only thing that takes time is learning how to develop from the start for ALL browsers. Once you learn WHY IE sucks, it's not that hard to develop a successful project without much extra development time.

Answer 5

Please stop developing for Internet Explorer 6 because IE6 Must Die for the Web to Move On.

After all YouTube is ditching support for it soon. And they're not the only one.

Answer 6

If it's a personal site, do whatever pleases you.

If it's a business site, consider the financial impact of each decision, comparing revenue lost from visitors on unsupported browsers compared to the development costs of supporting them. First gather browsers statististics: if it's an existing site, you should have them already. If it's a new site, you'll need to look for statistics from similar sites.

Of course, you don't need an all-or-nothing approach. You could support older versions of IE, but with limited functionality. Decide the maximum amount of resources you want to devote to IE6 and see what you can do under that restraint. You could also try grouping IE6 into the same category as mobile devices and other user agents with limited standards support. Not only do you not have to drop IE6 support, you've got even more browsers covered. I'm not saying this will be easy, or even possible, since IE6's behavior may be too different from the other low-web-tech user agents.

Answer 7

If Microsoft is urging users to ditch IE6, then should you really be spending time to accommodate those users?

I say build your site so that it's functional in IE 6 and pretty in everything else. Spending time on the latter for IE6 is way toO expensive to be worthwhile.

Answer 8

If support for IE6 is really needed (which is actually rarely the case), it's sufficient when a web page/application is usable with IE6.

If a user can access the information with his stone age browser, that's enough. You don't need to ensure that your css layout and fancy javascript stuff works in IE6. The people who still use IE 6 mostly won't even care about such things.

My personal strategy for IE 6 is: If the client does not explicitly want full IE 6 support, then he gets only the most basic IE 6 support. If he want's full support, he has to pay for the extra effort (which is significant ...).

Answer 9

It's not that hard to create sites that work in IE6. The only thing you have to do is

1. Don't use any CSS that's not CSS1
2. Use jQuery
3. Use IE6 as your development browser, and
4. Occasionally test with Firefox.

As long as you stick to those rules, your site is very likely to work with all browsers. You will run into random IE6 bugs sometimes that might take you an hour or two to work around, but it's usually not too bad.

Answer 10

I don't think there is a single good reason to even attempt to support IE6 unless your single biggest client is paying you huge sums of money to support it. In this case its a no brainer.

Versions of IE later than 6 have a standards mode, which helps substantially in getting most pages that look good in other browsers to look pretty ok in IE too.

Now look at the actual usage statistics for browsers and the different IE versions. Oddball browsers, which would be IE 6 and earlier, as well as the 'other' browsers that could be pretty much anything, represent 24ish percent of the sampled traffic. That's a pretty healthy chunk, but it might not reflect your actual users.

You can make good use of your logs for this. Have a look at the number of unique IPs and the browser each of those unique IP's is using. This can give you a good idea of what your users are actually using. Then take a look at total page views by the different IPs and compare the distributions according to browser. Probably some browsers are going to show more page views than others. It could be a sign when there is a sharp difference between the page views of two common user agents that one browser has some sort of usability shortcoming preventing those users from getting the higher number of page views

This is a perfect opportunity to use a profile directed optimization. You have the data that tells you that a hot spot (popular browser) is suffering in performance (fewer average pageviews), so you can look at working on just that specific browser.

Answer 11

The real question, to me, is

How do you know not to develop for IE5.5? FireFox 2? Lynx?

The answer to that question will help guide you as to how you can make a decision with IE6.

If you want to get really serious, you would look at similar sites' statistics, etc. On the other hand, spending a lot less time and effort, you can read some statistics like these in under two minutes, and then you take a mad guess as to:

1. whether your users are like the general statistic in their ie6 usage and
2. whether you care about whatever percentage you've decided is right... losing 10% of market would be worth it? 20%? 5%?

Or, least cost and probably equal effectiveness: you can flip a coin. IE6 is still just about as popular as IE7 or IE8, so if you leave it out, but then you're choosing less work over more accessibility. Which might be a very good choice.