March 17 2008

IE6, IE7, IE8, web standards

There’s been a lot of discussion over Internet Explorer 8’s handling of web standards. The main concern is whether or not version 8 should support web standards out of the box, or default to IE7’s less strict web standards behavior. Microsoft has gone as far as to make a decision to support IE7’s less strict web standards by default, only to announce that they were wrong and launched IE8 Beta in strict standards mode.

Articles regarding the issue:

• Compatibility and IE8 - Original announcement that IE8 will ship in IE7’s less strict web standards mode
• Beyond DOCTYPE: Web Standards, Forward Compatibility, and IE8 - A List Apart, making the announcement.
• The Internet Explorer lock-in, <META HTTP-EQUIV=”X-BALL-CHAIN”> - Community doesn’t like this announcement
• IE8 Reverses decision - Microsoft announces that they were wrong, and they WILL support strict standards by default
• Dear Computerworld, there’s a reason why IE8 breaks the web - Blaming broken sites on IE CSS hacks
• Martian Headsets - Vista’s decisions to not support old software, and implications of doing the same thing for IE8

First you’ll have to determine what authors mean when they say “web standards”. Are we talking simply about any version of HTML that validates, in addition to valid CSS? The discussions I’ve read don’t say, specifically, so I can only assume that’s what they mean. What they do all seem to agree on is that if I create CSS that takes advantage of improper CSS rendering for any version of Internet Explorer, those will now render properly in an web standards compliant IE8; making them useless. This I can understand.

For quite some time now we’ve been dealing with IE6, and a future that includes IE6 will require IE6 specific CSS to make things look right. So we continue doing what we need to do, easy enough. But how many of you have been hacking things specifically for IE7, in the short time it’s been available? My guess is not many. Like me, there might be a couple declarations you overwrite for IE7 in an IE only stylesheet, but surely nothing that will keep your site from functioning properly. This results in a stylesheet that every single version of IE reads, but is really filled with IE6 only hacks.

If you do this and test in a browser that already passes the Acid 2 test (Opera 9, Safari 3), as IE8 will, then you shouldn’t have many problems.

Example of adding a stylesheet that all versions of IE will use in your HTML doc:


<!--[if IE]>
<link rel="stylesheet" type="text/css" media="all" href="exploder.css" />
< ![endif]-->


In exploder.css put your hacks for both IE6 and IE7:

/* IE 7 & UP HACKS: */
.foo {padding-left: 7px;}

* .foo {padding-left: 14px;}


Since IE7 came out, the idea was to only hack for IE6, after all, what other browsers really need that much help?

When you hear “web standards” do you think about Javascript? Me either.

What I’m really missing in all of the talk about IE8 breaking sites is why these sites are actually breaking. Surely it’s not some silly CSS that was written specifically for IE7 that makes google maps and gmail fail, as the articles have lead me to believe. I’ve not seen other changes in IE 8 discussed, although I imagine they’ve occurred. Such as updates to the DOM and Javascript engine.

In gmail, the only modifications based on the users browser that I can see, are used in Javascript. Same goes for google maps.

When these articles refer to “web standards” are they including Javascript for some reason? When they mention that “sites have little hacks in them that only apply to IE” do they mean these hacks are more than just CSS hacks? Because if all the problems were related to “little CSS hacks that only apply to IE” then all this commotion would be a bit much, and the problems would be something any decent designer could handle; no?

interweb, tech